From efeb639ad7507ab43a8992333f730a43d6e7931c Mon Sep 17 00:00:00 2001
From: qnbs <155236708+qnbs@users.noreply.github.com>
Date: Wed, 12 Aug 2026 21:51:07 +0200
Subject: [PATCH 1/3] fix(ai): download progress bytes/speed + voice
 progress-scale bug (#333 item 1)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Voice model download (VoiceModelDownloadModal, transformers.js pipeline):
transformers.js's own onProgress payload is `{ progress: <0-100>, loaded:
<bytes>, total: <bytes> }` (verified against its readResponse() source —
progress = loaded/total*100). voiceCommandService.ts treated `progress` as if
it were already a 0-1 fraction, so the existing `Math.min(0.95, pct)` safety
clamp kicked in almost immediately (any progress value over 0.95%, i.e. after
the very first chunk) and stayed there for the rest of the download, making
the progress bar look stuck. Now derives the fraction from the payload's real
loaded/total byte counts (falling back to progress/100 only when absent), and
threads the real bytes through to the modal for a live "X MB of Y MB" + speed
display.

WebLLM text-model download (LocalAiDownloadProgress, services/localAiFacade):
the installed @mlc-ai/web-llm's own progress callback exposes only a 0-1
fraction, no structured byte counts. Adds a machine-readable per-model size
table (WEBLLM_MODEL_APPROX_MB, mirroring the free-text GB figures already in
WEBLLM_SUPPORTED_MODELS' labels) and derives an approximate downloaded/total
MB + speed from progress × that table — clearly labeled as an estimate, not
measured telemetry, since no real byte-level data is obtainable here without
patching library internals.

Both UIs share services/downloadProgressFormat.ts for MB/speed formatting.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
---
 CHANGELOG.md                                  |  14 +++
 README.md                                     |   8 +-
 .../settings/LocalAiDownloadProgress.tsx      |  27 ++++-
 components/voice/VoiceModelDownloadModal.tsx  |  43 +++++++
 locales/ar/settings.json                      |  30 ++---
 locales/de/settings.json                      |   4 +
 locales/el/settings.json                      |  30 ++---
 locales/en/settings.json                      |   4 +
 locales/es/settings.json                      |   4 +
 locales/eu/settings.json                      |  30 ++---
 locales/fa/settings.json                      |  30 ++---
 locales/fi/settings.json                      |  30 ++---
 locales/fr/settings.json                      |   4 +
 locales/he/settings.json                      |  30 ++---
 locales/hu/settings.json                      |  30 ++---
 locales/is/settings.json                      |  30 ++---
 locales/it/settings.json                      |   4 +
 locales/ja/settings.json                      |  30 ++---
 locales/ko/settings.json                      |  30 ++---
 locales/pt/settings.json                      |  30 ++---
 locales/ru/settings.json                      |  30 ++---
 locales/sv/settings.json                      |  30 ++---
 locales/zh/settings.json                      |  30 ++---
 packages/ai-core/src/index.ts                 |  15 +++
 public/locales/ar/bundle.json                 |  30 ++---
 public/locales/de/bundle.json                 |   4 +
 public/locales/el/bundle.json                 |  30 ++---
 public/locales/en/bundle.json                 |   4 +
 public/locales/es/bundle.json                 |   4 +
 public/locales/eu/bundle.json                 |  30 ++---
 public/locales/fa/bundle.json                 |  30 ++---
 public/locales/fi/bundle.json                 |  30 ++---
 public/locales/fr/bundle.json                 |   4 +
 public/locales/he/bundle.json                 |  30 ++---
 public/locales/hu/bundle.json                 |  30 ++---
 public/locales/is/bundle.json                 |  30 ++---
 public/locales/it/bundle.json                 |   4 +
 public/locales/ja/bundle.json                 |  30 ++---
 public/locales/ko/bundle.json                 |  30 ++---
 public/locales/pt/bundle.json                 |  30 ++---
 public/locales/ru/bundle.json                 |  30 ++---
 public/locales/sv/bundle.json                 |  30 ++---
 public/locales/zh/bundle.json                 |  30 ++---
 services/ai/inferenceProgressEmitter.ts       |  49 +++++++-
 services/downloadProgressFormat.ts            |  13 +++
 services/localAiFacade.ts                     |  11 +-
 services/voice/voiceCommandService.ts         |  20 +++-
 tests/unit/LocalAiDownloadProgress.test.tsx   | 110 +++++++++++++++++-
 .../voice/VoiceModelDownloadModal.test.tsx    |  43 ++++++-
 tests/unit/localAiFacade.test.ts              |   4 +-
 .../services/downloadProgressFormat.test.ts   |  29 +++++
 .../services/inferenceProgressEmitter.test.ts |  52 +++++++++
 .../voice/voiceCommandService.test.ts         |  86 ++++++++++++++
 types.ts                                      |   6 +
 54 files changed, 1023 insertions(+), 387 deletions(-)
 create mode 100644 services/downloadProgressFormat.ts
 create mode 100644 tests/unit/services/downloadProgressFormat.test.ts

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 4d3e99d2..c116aa97 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -26,6 +26,13 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   the race where an ordinary protected write could commit after a migration had already claimed
   ownership of the same store, which could otherwise land ciphertext under a superseded
   key/generation. (#339)
+- **Downloaded/total size and speed in both model-download progress UIs** (#333 item 1).
+  `VoiceModelDownloadModal` now shows real byte counts and transfer speed, sourced directly from
+  transformers.js's own progress payload (`loaded`/`total` fields it was already receiving but
+  discarding). `LocalAiDownloadProgress` (WebLLM text models) shows an **approximate** downloaded/
+  total MB and speed, derived from the existing 0-1 progress fraction × a new per-model known-size
+  table (`WEBLLM_MODEL_APPROX_MB`) — the installed `@mlc-ai/web-llm`'s own progress callback exposes
+  no structured byte counts, so this is clearly labeled as an estimate, not measured telemetry.
 
 ### Changed
 
@@ -60,6 +67,13 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - **Migration verification no longer re-scans already-verified stores on resume**, and a batch that
   reports progress without advancing its durable cursor is now rejected instead of being able to
   replay the same records indefinitely. (#337)
+- **Voice model download progress bar looked stuck at ~95% for most of the download.**
+  `voiceCommandService.ts`'s download progress handler treated transformers.js's `progress` field
+  as if it were already a 0-1 fraction; it's actually 0-100 (percent), so the existing
+  `Math.min(0.95, pct)` safety clamp kicked in almost immediately (any progress value over 0.95%,
+  i.e. after the very first chunk of the download) and stayed there until the download's final
+  "complete" dispatch. Now derives the fraction from the payload's real `loaded`/`total` byte
+  counts (falling back to `progress / 100` only when those are absent). (#333 item 1)
 
 ### Docs
 
diff --git a/README.md b/README.md
index 267e40fd..b5f890a0 100644
--- a/README.md
+++ b/README.md
@@ -12,7 +12,7 @@
   <img src="https://img.shields.io/badge/Version-v1.26.0-6366F1" alt="v1.26.0">
   <img src="https://img.shields.io/badge/Storage-IndexedDB_v8-F59E0B" alt="IndexedDB v8">
   <img src="https://img.shields.io/badge/PWA-v3.0-5BB974?logo=pwa" alt="PWA v3.0">
-  <img src="https://img.shields.io/badge/i18n-19_locales-2904_keys-0EA5E9" alt="i18n 19 locales — 2904 keys">
+  <img src="https://img.shields.io/badge/i18n-19_locales-2908_keys-0EA5E9" alt="i18n 19 locales — 2908 keys">
   <img src="https://img.shields.io/badge/Tests-6477%2B_%2F_532_files-22C55E" alt="6477+ tests / 532 files">
   <img src="https://img.shields.io/codecov/c/github/qnbs/WorldScript-Studio?logo=codecov&label=Coverage" alt="Codecov Coverage">
   <img src="https://img.shields.io/badge/License-MIT-22C55E" alt="License MIT">
@@ -397,7 +397,7 @@ Infrastructure-level features that keep the app fast and extensible as projects
 
 ### 🌐 Full Multi-Language Support
 
-Shipped UI locales with **2904 i18n keys** across all 19 languages — zero hardcoded user-facing strings:
+Shipped UI locales with **2908 i18n keys** across all 19 languages — zero hardcoded user-facing strings:
 
 - 🇩🇪 **German** (Deutsch)
 - 🇬🇧 **English**
@@ -506,7 +506,7 @@ The Settings → AI panel shows a live GPU status badge with adapter details and
 | **PDF Export**       | jsPDF                                                     | Client-side, configurable PDF document generation                    |
 | **Document Export**  | docx + jszip                                              | Word-compatible `.docx` generation (lazy-loaded)                     |
 | **PWA**              | Service Worker + Web App Manifest v3                     | Offline support, installability, Workbox chunking                    |
-| **i18n**             | Custom React Context (`I18nContext.tsx`)                  | 2904 keys × 19 locales (de/en/es/fr/it + ar/he/fa RTL Beta + ja/zh/pt/el/fi/sv/hu/is/eu/ru/ko Beta); EN fallback; `localStorage` persistence |
+| **i18n**             | Custom React Context (`I18nContext.tsx`)                  | 2908 keys × 19 locales (de/en/es/fr/it + ar/he/fa RTL Beta + ja/zh/pt/el/fi/sv/hu/is/eu/ru/ko Beta); EN fallback; `localStorage` persistence |
 | **Testing**          | Vitest 4.x (6477+ tests / 532 files) + Playwright E2E     | Unit/integration + cross-browser E2E; Stryker mutation (manual workflow)          |
 | **Code Quality**     | Biome (lint + format) + TypeScript 7 (tsgo) strict       | `--error-on-warnings` in CI; zero `any` policy                      |
 | **Visualization**    | Force-directed graph                                      | Interactive character relationship network                           |
@@ -708,7 +708,7 @@ The main pipeline is [`.github/workflows/ci.yml`](.github/workflows/ci.yml). Opt
 **Current test metrics (2026-07-30, CI-reported):**
 - **6477+ unit tests** across **532 test files** — all passing
 - Coverage thresholds: lines ≥ 74 · branches ≥ 60 · functions ≥ 67 · statements ≥ 72 — enforced in CI (see Codecov badge for live metrics)
-- i18n: **2904 keys × 19 locales** (en/de/fr/es/it + ar/he/fa RTL Beta + ja/zh/pt/el/fi/sv/hu/is/eu/ru/ko Beta)
+- i18n: **2908 keys × 19 locales** (en/de/fr/es/it + ar/he/fa RTL Beta + ja/zh/pt/el/fi/sv/hu/is/eu/ru/ko Beta)
 
 **CI-cloud-first workflow (recommended):** On constrained hardware run **`pnpm run lint && pnpm run i18n:check && pnpm run typecheck`** locally, then push and let CI handle coverage, E2E, Lighthouse, and Stryker. Authoritative numbers come from CI artifacts (Codecov, JUnit). After CI goes green, update the README badges and `AUDIT.md` quality-gate line from the reported metrics. See **[`docs/CI.md`](docs/CI.md) § Cloud CI-first vs local development** for the full post-merge doc-update checklist.
 
diff --git a/components/settings/LocalAiDownloadProgress.tsx b/components/settings/LocalAiDownloadProgress.tsx
index 2292f6b4..be941428 100644
--- a/components/settings/LocalAiDownloadProgress.tsx
+++ b/components/settings/LocalAiDownloadProgress.tsx
@@ -9,6 +9,7 @@ import {
   inferenceProgressEmitter,
   type WebLlmLoadProgress,
 } from '../../services/ai/inferenceProgressEmitter';
+import { formatMegabytes, formatMegabytesPerSecond } from '../../services/downloadProgressFormat';
 import { abortActivePreload, retryLastPreload } from '../../services/localAiFacade';
 
 export const LocalAiDownloadProgress: FC = () => {
@@ -46,6 +47,22 @@ export const LocalAiDownloadProgress: FC = () => {
     ? `${progressPct}% ${t<string>('settings.ai.localAi.downloadedLabel')}, ${etaText}`
     : `${progressPct}%`;
 
+  // QNBS-v3 (#333 item 1): approximate — see WebLlmLoadProgress's own doc comment. Only shown once
+  // both bounds are known (null whenever the active model id isn't in WEBLLM_MODEL_APPROX_MB).
+  const sizeText =
+    progress.loadedBytes != null && progress.totalBytes != null
+      ? t<string>('settings.ai.localAi.downloadSize', {
+          loaded: formatMegabytes(progress.loadedBytes),
+          total: formatMegabytes(progress.totalBytes),
+        })
+      : '';
+  const speedText =
+    progress.bytesPerSecond != null
+      ? t<string>('settings.ai.localAi.downloadSpeed', {
+          speed: formatMegabytesPerSecond(progress.bytesPerSecond),
+        })
+      : '';
+
   function handleCancel() {
     // QNBS-v3: actually abort the in-flight preload/worker task — not just hide the modal.
     abortActivePreload();
@@ -138,10 +155,18 @@ export const LocalAiDownloadProgress: FC = () => {
               />
             </div>
 
-            <div className="mb-4 flex justify-between text-xs text-[var(--sc-text-muted)]">
+            <div
+              className={`flex justify-between text-xs text-[var(--sc-text-muted)] ${sizeText || speedText ? 'mb-1' : 'mb-4'}`}
+            >
               <span>{progressPct}%</span>
               {etaText && <span>{etaText}</span>}
             </div>
+            {(sizeText || speedText) && (
+              <div className="mb-4 flex justify-between text-xs text-[var(--sc-text-muted)]">
+                <span>{sizeText}</span>
+                {speedText && <span>{speedText}</span>}
+              </div>
+            )}
 
             <button
               ref={cancelRef}
diff --git a/components/voice/VoiceModelDownloadModal.tsx b/components/voice/VoiceModelDownloadModal.tsx
index 20298ea5..2d9e069c 100644
--- a/components/voice/VoiceModelDownloadModal.tsx
+++ b/components/voice/VoiceModelDownloadModal.tsx
@@ -7,6 +7,7 @@ import React, { useCallback, useEffect, useRef, useState } from 'react';
 import { useAppDispatch, useAppSelector } from '../../app/hooks';
 import { settingsActions } from '../../features/settings/settingsSlice';
 import { useTranslation } from '../../hooks/useTranslation';
+import { formatMegabytes, formatMegabytesPerSecond } from '../../services/downloadProgressFormat';
 import { Button } from '../ui/Button';
 import { Modal } from '../ui/Modal';
 import { Progress } from '../ui/Progress';
@@ -31,10 +32,32 @@ export const VoiceModelDownloadModal = React.memo(function VoiceModelDownloadMod
   const { t } = useTranslation();
   const dispatch = useAppDispatch();
   const progress = useAppSelector((s) => s.settings.voice.wasmModelDownloadProgress ?? 0);
+  // QNBS-v3 (#333 item 1): real bytes from transformers.js's own progress payload (not an
+  // approximation — see wasmModelDownloadLoadedBytes's doc comment in types.ts).
+  const loadedBytes = useAppSelector((s) => s.settings.voice.wasmModelDownloadLoadedBytes);
+  const totalBytes = useAppSelector((s) => s.settings.voice.wasmModelDownloadTotalBytes);
   const [isDownloading, setIsDownloading] = useState(false);
   const [error, setError] = useState<string | null>(null);
   // QNBS-v3: CodeAnt P2-3 — AbortController lets cancel button stop an in-flight download
   const abortRef = useRef<AbortController | null>(null);
+  // QNBS-v3 (#333 item 1): wall-clock speed, computed client-side from real loaded-bytes deltas —
+  // the service layer only reports cumulative bytes, not a rate.
+  const downloadStartMsRef = useRef<number | null>(null);
+  const [bytesPerSecond, setBytesPerSecond] = useState<number | null>(null);
+
+  useEffect(() => {
+    if (!isDownloading) {
+      downloadStartMsRef.current = null;
+      setBytesPerSecond(null);
+      return;
+    }
+    if (downloadStartMsRef.current == null) {
+      downloadStartMsRef.current = Date.now();
+    }
+    if (loadedBytes == null) return;
+    const elapsedSeconds = (Date.now() - downloadStartMsRef.current) / 1000;
+    setBytesPerSecond(elapsedSeconds > 0.5 ? loadedBytes / elapsedSeconds : null);
+  }, [isDownloading, loadedBytes]);
 
   const handleDownload = useCallback(async () => {
     const ctrl = new AbortController();
@@ -87,6 +110,20 @@ export const VoiceModelDownloadModal = React.memo(function VoiceModelDownloadMod
   const modelName = modelType === 'stt' ? 'Whisper (STT)' : 'Kokoro (TTS)';
   const modelSize = modelType === 'stt' ? MODEL_SIZES.whisper : MODEL_SIZES.kokoro;
 
+  // QNBS-v3 (#333 item 1): real byte counts once the first progress tick with them has arrived;
+  // MODEL_SIZES above is only the pre-download estimate shown in the description text.
+  const sizeText =
+    loadedBytes != null && totalBytes != null
+      ? t('voice.modelDownload.progressBytes', {
+          loaded: formatMegabytes(loadedBytes),
+          total: formatMegabytes(totalBytes),
+        })
+      : '';
+  const speedText =
+    bytesPerSecond != null
+      ? t('voice.modelDownload.speed', { speed: formatMegabytesPerSecond(bytesPerSecond) })
+      : '';
+
   return (
     <Modal
       isOpen={isOpen}
@@ -109,6 +146,12 @@ export const VoiceModelDownloadModal = React.memo(function VoiceModelDownloadMod
             <p className="text-xs text-[var(--sc-text-tertiary)]" aria-live="polite">
               {t('voice.modelDownload.progress', { percent: String(Math.round(progress * 100)) })}
             </p>
+            {(sizeText || speedText) && (
+              <p className="flex justify-between text-xs text-[var(--sc-text-tertiary)]">
+                <span>{sizeText}</span>
+                {speedText && <span>{speedText}</span>}
+              </p>
+            )}
           </>
         )}
 
diff --git a/locales/ar/settings.json b/locales/ar/settings.json
index 5e4fa3bc..cb6243ec 100644
--- a/locales/ar/settings.json
+++ b/locales/ar/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "أُلغي التنزيل",
   "settings.ai.localAi.downloadEta": "يتبقّى ~{{seconds}} ثانية",
   "settings.ai.localAi.downloadFailed": "Download failed. Please try again.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "جارٍ تنزيل نموذج الذكاء الاصطناعي…",
   "settings.ai.localAi.downloadedLabel": "مُنزَّل",
   "settings.ai.localAi.downloadingButton": "Downloading…",
@@ -198,6 +200,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Not available",
   "settings.ai.localBackendPreset": "إعداد مسبق للخلفية المحلية",
+  "settings.ai.localDiagnostic.browserFetch": "طلب المتصفح",
+  "settings.ai.localDiagnostic.endpoint": "نقطة النهاية",
+  "settings.ai.localDiagnostic.models": "النماذج",
+  "settings.ai.localDiagnostic.tauriHttp": "HTTP الأصلي من Tauri",
+  "settings.ai.localDiagnostic.title": "تشخيص الاتصال",
+  "settings.ai.localDiagnostic.transport": "النقل",
   "settings.ai.ollamaBrowserNote": "‏Ollama متاح فقط في تطبيق سطح المكتب. تمنع سياسة أمان المحتوى للمتصفّح الاتصالات المباشرة بـ localhost.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -229,15 +237,9 @@
   "settings.ai.providerStatusConnected": "متصل",
   "settings.ai.providerStatusDisconnected": "غير متاح",
   "settings.ai.providerStatusLabel": "حالة الاتصال",
-  "settings.ai.providerStatusReady": "جاهز",
   "settings.ai.providerStatusNotTested": "لم يتم الاختبار",
+  "settings.ai.providerStatusReady": "جاهز",
   "settings.ai.providerStatusTesting": "جارٍ الاختبار…",
-  "settings.ai.localDiagnostic.title": "تشخيص الاتصال",
-  "settings.ai.localDiagnostic.endpoint": "نقطة النهاية",
-  "settings.ai.localDiagnostic.transport": "النقل",
-  "settings.ai.localDiagnostic.models": "النماذج",
-  "settings.ai.localDiagnostic.tauriHttp": "HTTP الأصلي من Tauri",
-  "settings.ai.localDiagnostic.browserFetch": "طلب المتصفح",
   "settings.ai.providerStatusUnavailableBrowser": "غير متوفر في المتصفح",
   "settings.ai.providerTitle": "مزوّد الذكاء الاصطناعي",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -260,8 +262,8 @@
   "settings.ai.testError.desktopRequired": "تتوفر Ollama والخوادم المحلية المتوافقة مع OpenAI فقط في تطبيق سطح المكتب. تحظر المتصفحات الاتصالات المباشرة من صفحات الويب بـ localhost (CORS وPrivate Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "لم يتم تعيين مفتاح API لـ {{provider}}",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU غير متوفر في هذا المتصفح — يحتاج WebLLM إلى WebGPU (جرّب Chrome/Edge أو فعّل العلامات).",
   "settings.ai.testError.pluginUnavailable": "شبكة الخادم المحلي غير متوفرة: فشل تحميل مكوّن HTTP لسطح المكتب.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -345,14 +347,14 @@
   "settings.appearance.presetDefault": "افتراضي",
   "settings.appearance.presetHint": "تتراكب الإعدادات المسبقة مع وضع الإضاءة. سيبيا الكاتب يتكيف مع كلا الوضعين.",
   "settings.appearance.presetSepia": "سيبيا الكاتب",
-  "settings.appearance.writingSurface": "سطح الكتابة",
-  "settings.appearance.writingSurfaceHint": "اختر نمط الخلفية الأنسب للكتابة الطويلة.",
-  "settings.appearance.writingSurfaceTextured": "منسوج",
-  "settings.appearance.writingSurfacePlain": "سادة",
   "settings.appearance.primaryColor": "اللون الأساسي",
   "settings.appearance.secondaryColor": "اللون الثانوي",
   "settings.appearance.theme": "سمة التطبيق",
   "settings.appearance.title": "المظهر",
+  "settings.appearance.writingSurface": "سطح الكتابة",
+  "settings.appearance.writingSurfaceHint": "اختر نمط الخلفية الأنسب للكتابة الطويلة.",
+  "settings.appearance.writingSurfacePlain": "سادة",
+  "settings.appearance.writingSurfaceTextured": "منسوج",
   "settings.categories.about": "حول",
   "settings.categories.accessibility": "إمكانية الوصول",
   "settings.categories.advancedAi": "ذكاء اصطناعي متقدّم",
@@ -689,13 +691,13 @@
   "settings.privacy.encryptionSetAction": "تعيين عبارة المرور",
   "settings.privacy.encryptionSetButton": "تفعيل وتشفير",
   "settings.privacy.encryptionSetup": "عيّن عبارة مرور لتشفير جميع المخطوطات واللقطات والإعدادات المُخزَّنة على هذا الجهاز.",
+  "settings.privacy.encryptionSetupFailed": "تعذر حفظ إعداد التشفير — تحقق من أن هذا المتصفح يسمح بالتخزين المحلي، ثم حاول مرة أخرى",
   "settings.privacy.encryptionTooManyAttempts": "محاولات فاشلة كثيرة جدًا. يرجى الانتظار {{seconds}} ثانية.",
   "settings.privacy.encryptionTooShort": "يجب ألا تقل عبارة المرور عن 8 أحرف",
   "settings.privacy.encryptionUnlockAction": "فتح التخزين",
   "settings.privacy.encryptionUnlockButton": "فتح",
   "settings.privacy.encryptionWarning": "إذا نسيت عبارة المرور، فلا يمكن استرجاع بياناتك.",
   "settings.privacy.encryptionWrongPassphrase": "عبارة مرور خاطئة — فشل فك التشفير",
-  "settings.privacy.encryptionSetupFailed": "تعذر حفظ إعداد التشفير — تحقق من أن هذا المتصفح يسمح بالتخزين المحلي، ثم حاول مرة أخرى",
   "settings.privacy.euDataResidency": "إقامة البيانات في الاتحاد الأوروبي",
   "settings.privacy.euDataResidencyHint": "EU Data Residency — keep all data within EU servers where applicable.",
   "settings.privacy.localStorageOnly": "التخزين المحلي فقط",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "جاري تحميل نموذج {{model}} ({{size}} ميغابايت). يُمكّن ذلك المعالجة الصوتية المحلية دون إرسال إلى السحابة.",
   "voice.modelDownload.error": "فشل التحميل: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% مكتمل",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "إعادة المحاولة",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "تحميل نموذج الصوت"
 }
diff --git a/locales/de/settings.json b/locales/de/settings.json
index 05662d24..c7780d36 100644
--- a/locales/de/settings.json
+++ b/locales/de/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "Download abgebrochen",
   "settings.ai.localAi.downloadEta": "ca. {{seconds}} Sek. verbleibend",
   "settings.ai.localAi.downloadFailed": "Download fehlgeschlagen. Bitte versuche es erneut.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB von {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "KI-Modell wird heruntergeladen…",
   "settings.ai.localAi.downloadedLabel": "heruntergeladen",
   "settings.ai.localAi.downloadingButton": "Wird heruntergeladen…",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "Lade {{model}}-Modell ({{size}} MB) herunter. Dies ermöglicht lokale Sprachverarbeitung ohne Cloud-Routing.",
   "voice.modelDownload.error": "Download fehlgeschlagen: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% abgeschlossen",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB von {{total}} MB",
   "voice.modelDownload.retry": "Wiederholen",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Sprachmodell-Download"
 }
diff --git a/locales/el/settings.json b/locales/el/settings.json
index 63b2ecd2..ade079b0 100644
--- a/locales/el/settings.json
+++ b/locales/el/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "Η λήψη ακυρώθηκε",
   "settings.ai.localAi.downloadEta": "περίπου. Απομένουν {{seconds}}s",
   "settings.ai.localAi.downloadFailed": "Η λήψη απέτυχε. Δοκιμάστε ξανά.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Λήψη μοντέλου AI…",
   "settings.ai.localAi.downloadedLabel": "κατεβάσει",
   "settings.ai.localAi.downloadingButton": "Λήψη…",
@@ -198,6 +200,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Μη διαθέσιμο",
   "settings.ai.localBackendPreset": "Τοπική προεπιλογή υποστήριξης",
+  "settings.ai.localDiagnostic.browserFetch": "Αίτημα προγράμματος περιήγησης",
+  "settings.ai.localDiagnostic.endpoint": "Τελικό σημείο",
+  "settings.ai.localDiagnostic.models": "Μοντέλα",
+  "settings.ai.localDiagnostic.tauriHttp": "Εγγενές HTTP Tauri",
+  "settings.ai.localDiagnostic.title": "Διαγνωστικά σύνδεσης",
+  "settings.ai.localDiagnostic.transport": "Μεταφορά",
   "settings.ai.ollamaBrowserNote": "Το Olama είναι διαθέσιμο μόνο στην εφαρμογή για υπολογιστές. Η Πολιτική Ασφάλειας Περιεχομένου του προγράμματος περιήγησης αποκλείει τις άμεσες συνδέσεις με τον localhost.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -229,15 +237,9 @@
   "settings.ai.providerStatusConnected": "Συνδεδεμένος",
   "settings.ai.providerStatusDisconnected": "Δυσεύρετος",
   "settings.ai.providerStatusLabel": "Κατάσταση σύνδεσης",
-  "settings.ai.providerStatusReady": "Ετοιμος",
   "settings.ai.providerStatusNotTested": "Δεν έχει δοκιμαστεί",
+  "settings.ai.providerStatusReady": "Ετοιμος",
   "settings.ai.providerStatusTesting": "Δοκιμή σε εξέλιξη…",
-  "settings.ai.localDiagnostic.title": "Διαγνωστικά σύνδεσης",
-  "settings.ai.localDiagnostic.endpoint": "Τελικό σημείο",
-  "settings.ai.localDiagnostic.transport": "Μεταφορά",
-  "settings.ai.localDiagnostic.models": "Μοντέλα",
-  "settings.ai.localDiagnostic.tauriHttp": "Εγγενές HTTP Tauri",
-  "settings.ai.localDiagnostic.browserFetch": "Αίτημα προγράμματος περιήγησης",
   "settings.ai.providerStatusUnavailableBrowser": "Μη διαθέσιμο στο πρόγραμμα περιήγησης",
   "settings.ai.providerTitle": "Πάροχος AI",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -260,8 +262,8 @@
   "settings.ai.testError.desktopRequired": "Το Ollama και οι τοπικοί διακομιστές συμβατοί με OpenAI είναι διαθέσιμοι μόνο στην εφαρμογή επιφάνειας εργασίας. Οι φυλλομετρητές αποκλείουν άμεσες συνδέσεις από ιστοσελίδες προς το localhost (CORS και Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "Δεν έχει οριστεί κλειδί API {{provider}}",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "Το WebGPU δεν είναι διαθέσιμο σε αυτό το πρόγραμμα περιήγησης — το WebLLM απαιτεί WebGPU (δοκιμάστε Chrome/Edge ή ενεργοποιήστε τις σημαίες).",
   "settings.ai.testError.pluginUnavailable": "Το δίκτυο τοπικού διακομιστή δεν είναι διαθέσιμο: η πρόσθετη λειτουργία HTTP επιφάνειας εργασίας απέτυχε να φορτωθεί.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -345,14 +347,14 @@
   "settings.appearance.presetDefault": "Αθέτηση",
   "settings.appearance.presetHint": "Οι προεπιλογές συνδυάζονται με την επιλογή λειτουργίας. Η Σέπια του συγγραφέα προσαρμόζεται πλήρως και στις δύο λειτουργίες.",
   "settings.appearance.presetSepia": "Σέπια του συγγραφέα",
-  "settings.appearance.writingSurface": "Writing surface",
-  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
-  "settings.appearance.writingSurfaceTextured": "Textured",
-  "settings.appearance.writingSurfacePlain": "Plain",
   "settings.appearance.primaryColor": "Πρωτεύον χρώμα",
   "settings.appearance.secondaryColor": "Δευτερεύον Χρώμα",
   "settings.appearance.theme": "Θέμα εφαρμογής",
   "settings.appearance.title": "Εμφάνιση",
+  "settings.appearance.writingSurface": "Writing surface",
+  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
+  "settings.appearance.writingSurfacePlain": "Plain",
+  "settings.appearance.writingSurfaceTextured": "Textured",
   "settings.categories.about": "Για",
   "settings.categories.accessibility": "Προσιτότητα",
   "settings.categories.advancedAi": "Προηγμένη τεχνητή νοημοσύνη",
@@ -689,13 +691,13 @@
   "settings.privacy.encryptionSetAction": "Ορίστε τη φράση πρόσβασης",
   "settings.privacy.encryptionSetButton": "Ενεργοποίηση & Κρυπτογράφηση",
   "settings.privacy.encryptionSetup": "Ορίστε μια φράση πρόσβασης για να κρυπτογραφήσετε όλα τα χειρόγραφα, τα στιγμιότυπα και τις ρυθμίσεις που είναι αποθηκευμένα σε αυτήν τη συσκευή.",
+  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.encryptionTooManyAttempts": "Πάρα πολλές αποτυχημένες προσπάθειες. Περιμένετε {{seconds}} δευτερόλεπτα.",
   "settings.privacy.encryptionTooShort": "Η φράση πρόσβασης πρέπει να αποτελείται από τουλάχιστον 8 χαρακτήρες",
   "settings.privacy.encryptionUnlockAction": "Ξεκλείδωμα αποθήκευσης",
   "settings.privacy.encryptionUnlockButton": "Ξεκλειδώνω",
   "settings.privacy.encryptionWarning": "Εάν ξεχάσετε τη φράση πρόσβασης, τα δεδομένα σας δεν μπορούν να ανακτηθούν.",
   "settings.privacy.encryptionWrongPassphrase": "Λανθασμένη φράση πρόσβασης — η αποκρυπτογράφηση απέτυχε",
-  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.euDataResidency": "Εντοπισμός δεδομένων ΕΕ",
   "settings.privacy.euDataResidencyHint": "Διατήρηση όλων των δεδομένων σε servers της Ευρωπαϊκής Ένωσης όπου είναι δυνατόν.",
   "settings.privacy.localStorageOnly": "Μόνο τοπικός χώρος αποθήκευσης",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "Λήψη μοντέλου {{model}} ({{size}} MB). Αυτό επιτρέπει την τοπική επεξεργασία φωνής χωρίς δρομολόγηση cloud.",
   "voice.modelDownload.error": "Η λήψη απέτυχε: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% ολοκληρώθηκε",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Δοκιμάζω πάλι",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Λήψη φωνητικού μοντέλου"
 }
diff --git a/locales/en/settings.json b/locales/en/settings.json
index 47b3c718..f8c535bd 100644
--- a/locales/en/settings.json
+++ b/locales/en/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "Download cancelled",
   "settings.ai.localAi.downloadEta": "approx. {{seconds}}s remaining",
   "settings.ai.localAi.downloadFailed": "Download failed. Please try again.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Downloading AI model…",
   "settings.ai.localAi.downloadedLabel": "downloaded",
   "settings.ai.localAi.downloadingButton": "Downloading…",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "Downloading {{model}} model ({{size}} MB). This enables local voice processing without cloud routing.",
   "voice.modelDownload.error": "Download failed: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% complete",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Retry",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Voice Model Download"
 }
diff --git a/locales/es/settings.json b/locales/es/settings.json
index 16d49970..d26fdebd 100644
--- a/locales/es/settings.json
+++ b/locales/es/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "Descarga cancelada",
   "settings.ai.localAi.downloadEta": "aprox. {{seconds}} s restantes",
   "settings.ai.localAi.downloadFailed": "La descarga falló. Inténtalo de nuevo.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB de {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Descargando modelo de IA…",
   "settings.ai.localAi.downloadedLabel": "descargado",
   "settings.ai.localAi.downloadingButton": "Descargando…",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "Descargando modelo {{model}} ({{size}} MB). Esto habilita el procesamiento de voz local sin enrutamiento en la nube.",
   "voice.modelDownload.error": "Descarga fallida: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% completado",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB de {{total}} MB",
   "voice.modelDownload.retry": "Reintentar",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Descarga de modelo de voz"
 }
diff --git a/locales/eu/settings.json b/locales/eu/settings.json
index bb8fd487..da932878 100644
--- a/locales/eu/settings.json
+++ b/locales/eu/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "Deskarga bertan behera utzi da",
   "settings.ai.localAi.downloadEta": "gutxi gorabehera. {{seconds}}s geratzen dira",
   "settings.ai.localAi.downloadFailed": "Deskargatu huts egin du. Mesedez, saiatu berriro.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "AI eredua deskargatzen…",
   "settings.ai.localAi.downloadedLabel": "deskargatu",
   "settings.ai.localAi.downloadingButton": "Deskargatzen…",
@@ -198,6 +200,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Ez dago eskuragarri",
   "settings.ai.localBackendPreset": "Backend lokala aurrez ezarri",
+  "settings.ai.localDiagnostic.browserFetch": "Arakatzailearen eskaera",
+  "settings.ai.localDiagnostic.endpoint": "Amaiera-puntua",
+  "settings.ai.localDiagnostic.models": "Ereduak",
+  "settings.ai.localDiagnostic.tauriHttp": "Tauriren jatorrizko HTTPa",
+  "settings.ai.localDiagnostic.title": "Konexio-diagnostikoa",
+  "settings.ai.localDiagnostic.transport": "Garraioa",
   "settings.ai.ollamaBrowserNote": "Ollama mahaigaineko aplikazioan bakarrik dago eskuragarri. Arakatzailearen Edukiaren Segurtasun-Politikak lokaleko ostalari konexio zuzenak blokeatzen ditu.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -229,15 +237,9 @@
   "settings.ai.providerStatusConnected": "Konektatuta",
   "settings.ai.providerStatusDisconnected": "Ez dago erabilgarri",
   "settings.ai.providerStatusLabel": "Konexioaren egoera",
-  "settings.ai.providerStatusReady": "Prest",
   "settings.ai.providerStatusNotTested": "Probatu gabe",
+  "settings.ai.providerStatusReady": "Prest",
   "settings.ai.providerStatusTesting": "Probatzen…",
-  "settings.ai.localDiagnostic.title": "Konexio-diagnostikoa",
-  "settings.ai.localDiagnostic.endpoint": "Amaiera-puntua",
-  "settings.ai.localDiagnostic.transport": "Garraioa",
-  "settings.ai.localDiagnostic.models": "Ereduak",
-  "settings.ai.localDiagnostic.tauriHttp": "Tauriren jatorrizko HTTPa",
-  "settings.ai.localDiagnostic.browserFetch": "Arakatzailearen eskaera",
   "settings.ai.providerStatusUnavailableBrowser": "Ez dago erabilgarri nabigatzailean",
   "settings.ai.providerTitle": "AI hornitzailea",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -260,8 +262,8 @@
   "settings.ai.testError.desktopRequired": "Ollama eta OpenAI-rekin bateragarriak diren zerbitzari lokalak mahaigaineko aplikazioan bakarrik daude erabilgarri. Nabigatzaileek web orrialdeetatik localhost-era zuzeneko konexioak blokeatzen dituzte (CORS eta Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "Ez da {{provider}} API gakorik ezarri",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU ez dago erabilgarri nabigatzaile honetan — WebLLM-ek WebGPU behar du (probatu Chrome/Edge edo gaitu banderak).",
   "settings.ai.testError.pluginUnavailable": "Zerbitzari lokalaren sarea ez dago erabilgarri: mahaigaineko HTTP pluginak huts egin du kargatzean.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -345,14 +347,14 @@
   "settings.appearance.presetDefault": "Lehenetsia",
   "settings.appearance.presetHint": "Aurrez ezarri geruza zure aukera ilun/argian. Idazlearen Sepia guztiz egokitzen da bi moduetara.",
   "settings.appearance.presetSepia": "Idazlearen Sepia",
-  "settings.appearance.writingSurface": "Idazteko gainazala",
-  "settings.appearance.writingSurfaceHint": "Aukeratu testu luzeak idazteko erosoen zaizun atzeko planoaren estiloa.",
-  "settings.appearance.writingSurfaceTextured": "Testuratua",
-  "settings.appearance.writingSurfacePlain": "Lisoa",
   "settings.appearance.primaryColor": "Lehen kolorea",
   "settings.appearance.secondaryColor": "Bigarren mailako kolorea",
   "settings.appearance.theme": "Aplikazioaren gaia",
   "settings.appearance.title": "Itxura",
+  "settings.appearance.writingSurface": "Idazteko gainazala",
+  "settings.appearance.writingSurfaceHint": "Aukeratu testu luzeak idazteko erosoen zaizun atzeko planoaren estiloa.",
+  "settings.appearance.writingSurfacePlain": "Lisoa",
+  "settings.appearance.writingSurfaceTextured": "Testuratua",
   "settings.categories.about": "Buruz",
   "settings.categories.accessibility": "Irisgarritasuna",
   "settings.categories.advancedAi": "AI aurreratua",
@@ -689,13 +691,13 @@
   "settings.privacy.encryptionSetAction": "Ezarri pasaesaldia",
   "settings.privacy.encryptionSetButton": "Gaitu eta enkriptatu",
   "settings.privacy.encryptionSetup": "Ezarri pasaesaldi bat gailu honetan gordetako eskuizkribu, argazki eta ezarpen guztiak enkriptatzeko.",
+  "settings.privacy.encryptionSetupFailed": "Ezin izan da enkriptazioaren konfigurazioa gorde — egiaztatu nabigatzaile honek biltegiratze lokala onartzen duela, eta saiatu berriro",
   "settings.privacy.encryptionTooManyAttempts": "Porrot egindako saiakera gehiegi. Mesedez, itxaron {{seconds}} segundo.",
   "settings.privacy.encryptionTooShort": "Pasaesaldiek 8 karaktere izan behar dituzte gutxienez",
   "settings.privacy.encryptionUnlockAction": "Desblokeatu biltegia",
   "settings.privacy.encryptionUnlockButton": "Desblokeatu",
   "settings.privacy.encryptionWarning": "Pasaesaldia ahazten baduzu, ezin izango dira zure datuak berreskuratu.",
   "settings.privacy.encryptionWrongPassphrase": "Pasaesaldi okerra — deszifratzeak huts egin du",
-  "settings.privacy.encryptionSetupFailed": "Ezin izan da enkriptazioaren konfigurazioa gorde — egiaztatu nabigatzaile honek biltegiratze lokala onartzen duela, eta saiatu berriro",
   "settings.privacy.euDataResidency": "EBko Datuen Egoitza",
   "settings.privacy.euDataResidencyHint": "Gorde datu guztiak Europar Batasuneko zerbitzarietan, hala badagokio.",
   "settings.privacy.localStorageOnly": "Tokiko biltegiratzea soilik",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "{{model}} eredua deskargatzen ({{size}} MB). Honek ahots lokala prozesatzea ahalbidetzen du hodeiko bideratzerik gabe.",
   "voice.modelDownload.error": "Deskargatu huts egin du: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% osatuta",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Saiatu berriro",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Ahots eredua deskargatu"
 }
diff --git a/locales/fa/settings.json b/locales/fa/settings.json
index 2585b4af..6f87bc8d 100644
--- a/locales/fa/settings.json
+++ b/locales/fa/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "دانلود لغو شد",
   "settings.ai.localAi.downloadEta": "تقریبا {{seconds}} باقی مانده است",
   "settings.ai.localAi.downloadFailed": "دانلود انجام نشد. لطفا دوباره امتحان کنید.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "در حال دانلود مدل هوش مصنوعی…",
   "settings.ai.localAi.downloadedLabel": "دانلود شده است",
   "settings.ai.localAi.downloadingButton": "در حال دانلود…",
@@ -198,6 +200,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "در دسترس نیست",
   "settings.ai.localBackendPreset": "از پیش تنظیم شده باطن محلی",
+  "settings.ai.localDiagnostic.browserFetch": "دریافت مرورگر",
+  "settings.ai.localDiagnostic.endpoint": "نقطه پایانی",
+  "settings.ai.localDiagnostic.models": "مدل‌ها",
+  "settings.ai.localDiagnostic.tauriHttp": "HTTP بومی Tauri",
+  "settings.ai.localDiagnostic.title": "عیب‌یابی اتصال",
+  "settings.ai.localDiagnostic.transport": "انتقال",
   "settings.ai.ollamaBrowserNote": "Ollama فقط در برنامه دسکتاپ در دسترس است. خط مشی امنیت محتوای مرورگر اتصال مستقیم به لوکال هاست را مسدود می کند.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -229,15 +237,9 @@
   "settings.ai.providerStatusConnected": "متصل شد",
   "settings.ai.providerStatusDisconnected": "در دسترس نیست",
   "settings.ai.providerStatusLabel": "وضعیت اتصال",
-  "settings.ai.providerStatusReady": "آماده است",
   "settings.ai.providerStatusNotTested": "آزمایش نشده",
+  "settings.ai.providerStatusReady": "آماده است",
   "settings.ai.providerStatusTesting": "در حال آزمایش…",
-  "settings.ai.localDiagnostic.title": "عیب‌یابی اتصال",
-  "settings.ai.localDiagnostic.endpoint": "نقطه پایانی",
-  "settings.ai.localDiagnostic.transport": "انتقال",
-  "settings.ai.localDiagnostic.models": "مدل‌ها",
-  "settings.ai.localDiagnostic.tauriHttp": "HTTP بومی Tauri",
-  "settings.ai.localDiagnostic.browserFetch": "دریافت مرورگر",
   "settings.ai.providerStatusUnavailableBrowser": "در مرورگر در دسترس نیست",
   "settings.ai.providerTitle": "ارائه دهنده هوش مصنوعی",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -260,8 +262,8 @@
   "settings.ai.testError.desktopRequired": "Ollama و سرورهای محلی سازگار با OpenAI فقط در برنامه دسکتاپ در دسترس هستند. مرورگرها اتصالات مستقیم از صفحات وب به localhost را مسدود می‌کنند (CORS و Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "کلید API {{provider}} تنظیم نشده است",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU در این مرورگر در دسترس نیست — WebLLM به WebGPU نیاز دارد (Chrome/Edge را امتحان کنید یا پرچم‌ها را فعال کنید).",
   "settings.ai.testError.pluginUnavailable": "شبکه‌سازی سرور محلی در دسترس نیست: بارگذاری افزونه HTTP دسکتاپ ناموفق بود.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -345,14 +347,14 @@
   "settings.appearance.presetDefault": "پیش فرض",
   "settings.appearance.presetHint": "لایه از پیش تنظیم شده در انتخاب تاریک/روشن شما. Sepia نویسنده به طور کامل با هر دو حالت سازگار است.",
   "settings.appearance.presetSepia": "سپیای نویسنده",
-  "settings.appearance.writingSurface": "سطح نوشتن",
-  "settings.appearance.writingSurfaceHint": "سبک پس‌زمینه‌ای را که برای نوشتن طولانی راحت‌تر است انتخاب کنید.",
-  "settings.appearance.writingSurfaceTextured": "بافت‌دار",
-  "settings.appearance.writingSurfacePlain": "ساده",
   "settings.appearance.primaryColor": "رنگ اولیه",
   "settings.appearance.secondaryColor": "رنگ ثانویه",
   "settings.appearance.theme": "تم برنامه",
   "settings.appearance.title": "ظاهر",
+  "settings.appearance.writingSurface": "سطح نوشتن",
+  "settings.appearance.writingSurfaceHint": "سبک پس‌زمینه‌ای را که برای نوشتن طولانی راحت‌تر است انتخاب کنید.",
+  "settings.appearance.writingSurfacePlain": "ساده",
+  "settings.appearance.writingSurfaceTextured": "بافت‌دار",
   "settings.categories.about": "درباره",
   "settings.categories.accessibility": "قابلیت دسترسی",
   "settings.categories.advancedAi": "هوش مصنوعی پیشرفته",
@@ -689,13 +691,13 @@
   "settings.privacy.encryptionSetAction": "عبارت عبور را تنظیم کنید",
   "settings.privacy.encryptionSetButton": "فعال و رمزگذاری",
   "settings.privacy.encryptionSetup": "یک عبارت عبور برای رمزگذاری همه دست نوشته ها، عکس های فوری و تنظیمات ذخیره شده در این دستگاه تنظیم کنید.",
+  "settings.privacy.encryptionSetupFailed": "راه‌اندازی رمزگذاری ذخیره نشد — بررسی کنید که این مرورگر اجازه ذخیره‌سازی محلی را می‌دهد، سپس دوباره امتحان کنید",
   "settings.privacy.encryptionTooManyAttempts": "تلاش های ناموفق بسیار زیاد. لطفا {{seconds}} ثانیه صبر کنید.",
   "settings.privacy.encryptionTooShort": "عبارت عبور باید حداقل 8 کاراکتر باشد",
   "settings.privacy.encryptionUnlockAction": "قفل فضای ذخیره سازی را باز کنید",
   "settings.privacy.encryptionUnlockButton": "باز کردن قفل",
   "settings.privacy.encryptionWarning": "اگر رمز عبور خود را فراموش کنید، اطلاعات شما قابل بازیابی نیست.",
   "settings.privacy.encryptionWrongPassphrase": "عبارت عبور اشتباه - رمزگشایی انجام نشد",
-  "settings.privacy.encryptionSetupFailed": "راه‌اندازی رمزگذاری ذخیره نشد — بررسی کنید که این مرورگر اجازه ذخیره‌سازی محلی را می‌دهد، سپس دوباره امتحان کنید",
   "settings.privacy.euDataResidency": "اقامت داده اتحادیه اروپا",
   "settings.privacy.euDataResidencyHint": "در صورت لزوم، تمام داده ها را در سرورهای اتحادیه اروپا نگه دارید.",
   "settings.privacy.localStorageOnly": "فقط فضای ذخیره‌سازی محلی",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "در حال دانلود مدل {{model}} ({{size}} مگابایت). این امکان پردازش صدای محلی را بدون مسیریابی ابری فراهم می کند.",
   "voice.modelDownload.error": "دانلود انجام نشد: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% کامل شده است",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "دوباره امتحان کنید",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "دانلود مدل صوتی"
 }
diff --git a/locales/fi/settings.json b/locales/fi/settings.json
index 3ce432b2..c0e6813e 100644
--- a/locales/fi/settings.json
+++ b/locales/fi/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "Lataus peruutettu",
   "settings.ai.localAi.downloadEta": "noin {{seconds}}s jäljellä",
   "settings.ai.localAi.downloadFailed": "Lataus epäonnistui. Yritä uudelleen.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Ladataan tekoälymallia…",
   "settings.ai.localAi.downloadedLabel": "ladattu",
   "settings.ai.localAi.downloadingButton": "Ladataan…",
@@ -198,6 +200,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Ei saatavilla",
   "settings.ai.localBackendPreset": "Paikallinen taustajärjestelmän esiasetus",
+  "settings.ai.localDiagnostic.browserFetch": "Selaimen pyyntö",
+  "settings.ai.localDiagnostic.endpoint": "Päätepiste",
+  "settings.ai.localDiagnostic.models": "Mallit",
+  "settings.ai.localDiagnostic.tauriHttp": "Taurin natiivi HTTP",
+  "settings.ai.localDiagnostic.title": "Yhteysdiagnostiikka",
+  "settings.ai.localDiagnostic.transport": "Siirto",
   "settings.ai.ollamaBrowserNote": "Ollama on saatavilla vain työpöytäsovelluksessa. Selaimen sisällön suojauskäytäntö estää suorat yhteydet localhostiin.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -229,15 +237,9 @@
   "settings.ai.providerStatusConnected": "Yhdistetty",
   "settings.ai.providerStatusDisconnected": "Ei saatavilla",
   "settings.ai.providerStatusLabel": "Yhteyden tila",
-  "settings.ai.providerStatusReady": "Valmis",
   "settings.ai.providerStatusNotTested": "Ei testattu",
+  "settings.ai.providerStatusReady": "Valmis",
   "settings.ai.providerStatusTesting": "Testataan…",
-  "settings.ai.localDiagnostic.title": "Yhteysdiagnostiikka",
-  "settings.ai.localDiagnostic.endpoint": "Päätepiste",
-  "settings.ai.localDiagnostic.transport": "Siirto",
-  "settings.ai.localDiagnostic.models": "Mallit",
-  "settings.ai.localDiagnostic.tauriHttp": "Taurin natiivi HTTP",
-  "settings.ai.localDiagnostic.browserFetch": "Selaimen pyyntö",
   "settings.ai.providerStatusUnavailableBrowser": "Ei saatavilla selaimessa",
   "settings.ai.providerTitle": "AI Provider",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -260,8 +262,8 @@
   "settings.ai.testError.desktopRequired": "Ollama ja paikalliset OpenAI-yhteensopivat palvelimet ovat käytettävissä vain työpöytäsovelluksessa. Selaimet estävät suorat yhteydet verkkosivuilta osoitteeseen localhost (CORS ja Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "{{provider}}-API-avainta ei ole asetettu",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU ei ole käytettävissä tässä selaimessa — WebLLM tarvitsee WebGPU:n (kokeile Chromea/Edgeä tai ota liput käyttöön).",
   "settings.ai.testError.pluginUnavailable": "Paikallisen palvelimen verkko ei ole käytettävissä: työpöydän HTTP-lisäosan lataus epäonnistui.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -345,14 +347,14 @@
   "settings.appearance.presetDefault": "Oletus",
   "settings.appearance.presetHint": "Esiasettaa kerroksen tumman/vaalean valinnan mukaan. Writer's Sepia mukautuu täysin molempiin tiloihin.",
   "settings.appearance.presetSepia": "Kirjailijan Seepia",
-  "settings.appearance.writingSurface": "Writing surface",
-  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
-  "settings.appearance.writingSurfaceTextured": "Textured",
-  "settings.appearance.writingSurfacePlain": "Plain",
   "settings.appearance.primaryColor": "Ensisijainen väri",
   "settings.appearance.secondaryColor": "Toissijainen väri",
   "settings.appearance.theme": "Sovelluksen teema",
   "settings.appearance.title": "Ulkonäkö",
+  "settings.appearance.writingSurface": "Writing surface",
+  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
+  "settings.appearance.writingSurfacePlain": "Plain",
+  "settings.appearance.writingSurfaceTextured": "Textured",
   "settings.categories.about": "Noin",
   "settings.categories.accessibility": "Esteettömyys",
   "settings.categories.advancedAi": "Kehittynyt AI",
@@ -689,13 +691,13 @@
   "settings.privacy.encryptionSetAction": "Aseta salasana",
   "settings.privacy.encryptionSetButton": "Ota käyttöön ja salaa",
   "settings.privacy.encryptionSetup": "Aseta tunnuslause, joka salaa kaikki tähän laitteeseen tallennetut käsikirjoitukset, tilannevedokset ja asetukset.",
+  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.encryptionTooManyAttempts": "Liian monta epäonnistunutta yritystä. Odota {{seconds}} sekuntia.",
   "settings.privacy.encryptionTooShort": "Tunnuslauseessa on oltava vähintään 8 merkkiä",
   "settings.privacy.encryptionUnlockAction": "Avaa tallennustila",
   "settings.privacy.encryptionUnlockButton": "Avata",
   "settings.privacy.encryptionWarning": "Jos unohdat salasanasi, tietojasi ei voida palauttaa.",
   "settings.privacy.encryptionWrongPassphrase": "Väärä tunnuslause – salauksen purku epäonnistui",
-  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.euDataResidency": "EU Data Residency",
   "settings.privacy.euDataResidencyHint": "Säilytä kaikki tiedot tarvittaessa Euroopan unionin palvelimilla.",
   "settings.privacy.localStorageOnly": "Vain paikallinen tallennustila",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "Ladataan mallia {{model}} ({{size}} Mt). Tämä mahdollistaa paikallisen äänenkäsittelyn ilman pilvireititystä.",
   "voice.modelDownload.error": "Lataus epäonnistui: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% valmiina",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Yritä uudelleen",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Äänimallin lataus"
 }
diff --git a/locales/fr/settings.json b/locales/fr/settings.json
index f08d2242..2f7af0e7 100644
--- a/locales/fr/settings.json
+++ b/locales/fr/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "Téléchargement annulé",
   "settings.ai.localAi.downloadEta": "env. {{seconds}} s restantes",
   "settings.ai.localAi.downloadFailed": "Échec du téléchargement. Veuillez réessayer.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} Mo sur {{total}} Mo",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} Mo/s",
   "settings.ai.localAi.downloadTitle": "Téléchargement du modèle IA…",
   "settings.ai.localAi.downloadedLabel": "téléchargé",
   "settings.ai.localAi.downloadingButton": "Téléchargement…",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "Téléchargement du modèle {{model}} ({{size}} Mo). Cela permet le traitement vocal local sans routage cloud.",
   "voice.modelDownload.error": "Échec du téléchargement : {{error}}",
   "voice.modelDownload.progress": "{{percent}}% terminé",
+  "voice.modelDownload.progressBytes": "{{loaded}} Mo sur {{total}} Mo",
   "voice.modelDownload.retry": "Réessayer",
+  "voice.modelDownload.speed": "{{speed}} Mo/s",
   "voice.modelDownload.title": "Téléchargement du modèle vocal"
 }
diff --git a/locales/he/settings.json b/locales/he/settings.json
index 066abae7..c1df65d4 100644
--- a/locales/he/settings.json
+++ b/locales/he/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "ההורדה בוטלה",
   "settings.ai.localAi.downloadEta": "נותרו כ‑{{seconds}} שנ׳",
   "settings.ai.localAi.downloadFailed": "Download failed. Please try again.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "מוריד מודל AI…",
   "settings.ai.localAi.downloadedLabel": "הורד",
   "settings.ai.localAi.downloadingButton": "Downloading…",
@@ -198,6 +200,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Not available",
   "settings.ai.localBackendPreset": "הגדרה מוגדרת מראש של מנוע אחורי מקומי",
+  "settings.ai.localDiagnostic.browserFetch": "בקשת דפדפן",
+  "settings.ai.localDiagnostic.endpoint": "נקודת קצה",
+  "settings.ai.localDiagnostic.models": "מודלים",
+  "settings.ai.localDiagnostic.tauriHttp": "HTTP מקורי של Tauri",
+  "settings.ai.localDiagnostic.title": "אבחון חיבור",
+  "settings.ai.localDiagnostic.transport": "תעבורה",
   "settings.ai.ollamaBrowserNote": "‏Ollama זמין רק באפליקציית שולחן העבודה. מדיניות אבטחת התוכן של הדפדפן חוסמת חיבורים ישירים ל‑localhost.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -229,15 +237,9 @@
   "settings.ai.providerStatusConnected": "מחובר",
   "settings.ai.providerStatusDisconnected": "לא זמין",
   "settings.ai.providerStatusLabel": "סטטוס חיבור",
-  "settings.ai.providerStatusReady": "מוכן",
   "settings.ai.providerStatusNotTested": "לא נבדק",
+  "settings.ai.providerStatusReady": "מוכן",
   "settings.ai.providerStatusTesting": "בודק…",
-  "settings.ai.localDiagnostic.title": "אבחון חיבור",
-  "settings.ai.localDiagnostic.endpoint": "נקודת קצה",
-  "settings.ai.localDiagnostic.transport": "תעבורה",
-  "settings.ai.localDiagnostic.models": "מודלים",
-  "settings.ai.localDiagnostic.tauriHttp": "HTTP מקורי של Tauri",
-  "settings.ai.localDiagnostic.browserFetch": "בקשת דפדפן",
   "settings.ai.providerStatusUnavailableBrowser": "לא זמין בדפדפן",
   "settings.ai.providerTitle": "ספק AI",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -260,8 +262,8 @@
   "settings.ai.testError.desktopRequired": "Ollama ושרתים מקומיים תואמי OpenAI זמינים רק באפליקציית שולחן העבודה. דפדפנים חוסמים חיבורים ישירים מדפי אינטרנט אל localhost (CORS ו-Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "לא הוגדר מפתח API עבור {{provider}}",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU אינו זמין בדפדפן זה — WebLLM דורש WebGPU (נסה Chrome/Edge או הפעל דגלים).",
   "settings.ai.testError.pluginUnavailable": "רשת השרת המקומי אינה זמינה: תוסף ה-HTTP של שולחן העבודה נכשל בטעינה.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -345,14 +347,14 @@
   "settings.appearance.presetDefault": "ברירת מחדל",
   "settings.appearance.presetHint": "ההגדרות המוגדרות מראש מצטברות עם בחירת המצב. ספיה של הכותב מסתגל לשני המצבים.",
   "settings.appearance.presetSepia": "ספיה של הכותב",
-  "settings.appearance.writingSurface": "משטח כתיבה",
-  "settings.appearance.writingSurfaceHint": "בחרו את סגנון הרקע הנוח ביותר לכתיבה ארוכה.",
-  "settings.appearance.writingSurfaceTextured": "בעל מרקם",
-  "settings.appearance.writingSurfacePlain": "חלק",
   "settings.appearance.primaryColor": "צבע ראשי",
   "settings.appearance.secondaryColor": "צבע משני",
   "settings.appearance.theme": "ערכת נושא של האפליקציה",
   "settings.appearance.title": "מראה",
+  "settings.appearance.writingSurface": "משטח כתיבה",
+  "settings.appearance.writingSurfaceHint": "בחרו את סגנון הרקע הנוח ביותר לכתיבה ארוכה.",
+  "settings.appearance.writingSurfacePlain": "חלק",
+  "settings.appearance.writingSurfaceTextured": "בעל מרקם",
   "settings.categories.about": "אודות",
   "settings.categories.accessibility": "נגישות",
   "settings.categories.advancedAi": "AI מתקדם",
@@ -689,13 +691,13 @@
   "settings.privacy.encryptionSetAction": "הגדרת ביטוי סיסמה",
   "settings.privacy.encryptionSetButton": "הפעלה והצפנה",
   "settings.privacy.encryptionSetup": "הגדירו ביטוי סיסמה כדי להצפין את כל כתבי היד, תמונות המצב וההגדרות המאוחסנים במכשיר זה.",
+  "settings.privacy.encryptionSetupFailed": "לא ניתן היה לשמור את הגדרת ההצפנה — ודאו שדפדפן זה מאפשר אחסון מקומי ונסו שוב",
   "settings.privacy.encryptionTooManyAttempts": "יותר מדי ניסיונות כושלים. אנא המתינו {{seconds}} שניות.",
   "settings.privacy.encryptionTooShort": "ביטוי הסיסמה חייב להכיל לפחות 8 תווים",
   "settings.privacy.encryptionUnlockAction": "פתיחת אחסון",
   "settings.privacy.encryptionUnlockButton": "פתיחה",
   "settings.privacy.encryptionWarning": "אם תשכחו את ביטוי הסיסמה, לא ניתן יהיה לשחזר את הנתונים שלכם.",
   "settings.privacy.encryptionWrongPassphrase": "ביטוי סיסמה שגוי — הפענוח נכשל",
-  "settings.privacy.encryptionSetupFailed": "לא ניתן היה לשמור את הגדרת ההצפנה — ודאו שדפדפן זה מאפשר אחסון מקומי ונסו שוב",
   "settings.privacy.euDataResidency": "אחסון נתונים באיחוד האירופי",
   "settings.privacy.euDataResidencyHint": "EU Data Residency — keep all data within EU servers where applicable.",
   "settings.privacy.localStorageOnly": "אחסון מקומי בלבד",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "מוריד את מודל {{model}} ({{size}} מגהבייט). זה מאפשר עיבוד קול מקומי ללא ניתוב לענן.",
   "voice.modelDownload.error": "ההורדה נכשלה: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% הושלם",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "נסה שוב",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "הורדת מודל קול"
 }
diff --git a/locales/hu/settings.json b/locales/hu/settings.json
index 50b8f94f..1317d4ac 100644
--- a/locales/hu/settings.json
+++ b/locales/hu/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "Letöltés megszakítva",
   "settings.ai.localAi.downloadEta": "kb. {{seconds}}s van hátra",
   "settings.ai.localAi.downloadFailed": "Letöltés sikertelen. Kérjük, próbálja újra.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "AI-modell letöltése…",
   "settings.ai.localAi.downloadedLabel": "letöltve",
   "settings.ai.localAi.downloadingButton": "Letöltés…",
@@ -198,6 +200,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Nem elérhető",
   "settings.ai.localBackendPreset": "Helyi háttér előre beállított",
+  "settings.ai.localDiagnostic.browserFetch": "Böngészőlekérés",
+  "settings.ai.localDiagnostic.endpoint": "Végpont",
+  "settings.ai.localDiagnostic.models": "Modellek",
+  "settings.ai.localDiagnostic.tauriHttp": "Tauri natív HTTP",
+  "settings.ai.localDiagnostic.title": "Kapcsolati diagnosztika",
+  "settings.ai.localDiagnostic.transport": "Átvitel",
   "settings.ai.ollamaBrowserNote": "Az Ollama csak az asztali alkalmazásban érhető el. A böngésző tartalombiztonsági házirendje blokkolja a közvetlen kapcsolatot a localhosttal.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -229,15 +237,9 @@
   "settings.ai.providerStatusConnected": "Csatlakozva",
   "settings.ai.providerStatusDisconnected": "Nem elérhető",
   "settings.ai.providerStatusLabel": "Kapcsolat állapota",
-  "settings.ai.providerStatusReady": "Kész",
   "settings.ai.providerStatusNotTested": "Nincs tesztelve",
+  "settings.ai.providerStatusReady": "Kész",
   "settings.ai.providerStatusTesting": "Tesztelés…",
-  "settings.ai.localDiagnostic.title": "Kapcsolati diagnosztika",
-  "settings.ai.localDiagnostic.endpoint": "Végpont",
-  "settings.ai.localDiagnostic.transport": "Átvitel",
-  "settings.ai.localDiagnostic.models": "Modellek",
-  "settings.ai.localDiagnostic.tauriHttp": "Tauri natív HTTP",
-  "settings.ai.localDiagnostic.browserFetch": "Böngészőlekérés",
   "settings.ai.providerStatusUnavailableBrowser": "Böngészőben nem érhető el",
   "settings.ai.providerTitle": "AI szolgáltató",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -260,8 +262,8 @@
   "settings.ai.testError.desktopRequired": "Az Ollama és a helyi OpenAI-kompatibilis kiszolgálók csak az asztali alkalmazásban érhetők el. A böngészők blokkolják a weboldalak közvetlen kapcsolatait a localhosttal (CORS és Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "Nincs beállítva {{provider}} API-kulcs",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "A WebGPU nem érhető el ebben a böngészőben — a WebLLM-nek WebGPU-ra van szüksége (próbáld a Chrome-ot/Edge-et, vagy engedélyezd a jelzőket).",
   "settings.ai.testError.pluginUnavailable": "A helyi kiszolgálóhálózat nem érhető el: az asztali HTTP bővítmény betöltése sikertelen.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -345,14 +347,14 @@
   "settings.appearance.presetDefault": "Alapértelmezett",
   "settings.appearance.presetHint": "Előre beállítja a réteget a sötét/világos választás szerint. A Writer's Sepia teljes mértékben alkalmazkodik mindkét módhoz.",
   "settings.appearance.presetSepia": "Írói szépia",
-  "settings.appearance.writingSurface": "Writing surface",
-  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
-  "settings.appearance.writingSurfaceTextured": "Textured",
-  "settings.appearance.writingSurfacePlain": "Plain",
   "settings.appearance.primaryColor": "Elsődleges szín",
   "settings.appearance.secondaryColor": "Másodlagos szín",
   "settings.appearance.theme": "Alkalmazási téma",
   "settings.appearance.title": "Megjelenés",
+  "settings.appearance.writingSurface": "Writing surface",
+  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
+  "settings.appearance.writingSurfacePlain": "Plain",
+  "settings.appearance.writingSurfaceTextured": "Textured",
   "settings.categories.about": "Körülbelül",
   "settings.categories.accessibility": "Megközelíthetőség",
   "settings.categories.advancedAi": "Fejlett AI",
@@ -689,13 +691,13 @@
   "settings.privacy.encryptionSetAction": "Állítsa be a jelszót",
   "settings.privacy.encryptionSetButton": "Engedélyezés és titkosítás",
   "settings.privacy.encryptionSetup": "Állítson be egy jelszót az eszközön tárolt összes kézirat, pillanatkép és beállítás titkosításához.",
+  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.encryptionTooManyAttempts": "Túl sok sikertelen próbálkozás. Kérjük, várjon {{seconds}} másodpercet.",
   "settings.privacy.encryptionTooShort": "A jelszónak legalább 8 karakterből kell állnia",
   "settings.privacy.encryptionUnlockAction": "Tárolás feloldása",
   "settings.privacy.encryptionUnlockButton": "Kinyit",
   "settings.privacy.encryptionWarning": "Ha elfelejti jelszavát, adatait nem lehet visszaállítani.",
   "settings.privacy.encryptionWrongPassphrase": "Hibás jelszó – a visszafejtés nem sikerült",
-  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.euDataResidency": "EU Data Residency",
   "settings.privacy.euDataResidencyHint": "Adott esetben minden adatot tároljon az Európai Unió szerverein.",
   "settings.privacy.localStorageOnly": "Csak helyi tárhely",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "{{model}} modell letöltése ({{size}} MB). Ez lehetővé teszi a helyi hangfeldolgozást felhőalapú útválasztás nélkül.",
   "voice.modelDownload.error": "Sikertelen letöltés: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% kész",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Próbálja újra",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Hangmodell letöltése"
 }
diff --git a/locales/is/settings.json b/locales/is/settings.json
index 89d48005..27bb1817 100644
--- a/locales/is/settings.json
+++ b/locales/is/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "Hætt við niðurhal",
   "settings.ai.localAi.downloadEta": "ca. {{seconds}}s eftir",
   "settings.ai.localAi.downloadFailed": "Niðurhal mistókst. Vinsamlegast reyndu aftur.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Sækir gervigreind líkan…",
   "settings.ai.localAi.downloadedLabel": "niðurhalað",
   "settings.ai.localAi.downloadingButton": "Sækir…",
@@ -198,6 +200,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Ekki í boði",
   "settings.ai.localBackendPreset": "Forstilla staðbundin bakenda",
+  "settings.ai.localDiagnostic.browserFetch": "Vafrafyrirspurn",
+  "settings.ai.localDiagnostic.endpoint": "Endapunktur",
+  "settings.ai.localDiagnostic.models": "Líkön",
+  "settings.ai.localDiagnostic.tauriHttp": "Innbyggt Tauri HTTP",
+  "settings.ai.localDiagnostic.title": "Tengigreining",
+  "settings.ai.localDiagnostic.transport": "Flutningur",
   "settings.ai.ollamaBrowserNote": "Ollama er aðeins fáanlegt í skjáborðsforritinu. Innihaldsöryggisstefna vafrans lokar á beinar tengingar við localhost.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -229,15 +237,9 @@
   "settings.ai.providerStatusConnected": "Tengdur",
   "settings.ai.providerStatusDisconnected": "Ekki tiltækt",
   "settings.ai.providerStatusLabel": "Tengingarstaða",
-  "settings.ai.providerStatusReady": "Tilbúið",
   "settings.ai.providerStatusNotTested": "Ekki prófað",
+  "settings.ai.providerStatusReady": "Tilbúið",
   "settings.ai.providerStatusTesting": "Prófar…",
-  "settings.ai.localDiagnostic.title": "Tengigreining",
-  "settings.ai.localDiagnostic.endpoint": "Endapunktur",
-  "settings.ai.localDiagnostic.transport": "Flutningur",
-  "settings.ai.localDiagnostic.models": "Líkön",
-  "settings.ai.localDiagnostic.tauriHttp": "Innbyggt Tauri HTTP",
-  "settings.ai.localDiagnostic.browserFetch": "Vafrafyrirspurn",
   "settings.ai.providerStatusUnavailableBrowser": "Ekki í boði í vafra",
   "settings.ai.providerTitle": "AI veitandi",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -260,8 +262,8 @@
   "settings.ai.testError.desktopRequired": "Ollama og staðbundnir OpenAI-samhæfir netþjónar eru aðeins tiltækir í skjáborðsforritinu. Vafrar loka fyrir bein tengsl frá vefsíðum við localhost (CORS og Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "Enginn {{provider}} API lykill stilltur",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU er ekki tiltækt í þessum vafra — WebLLM þarfnast WebGPU (prófaðu Chrome/Edge eða virkjaðu flögg).",
   "settings.ai.testError.pluginUnavailable": "Staðbundið netþjónasamband ekki tiltækt: ekki tókst að hlaða HTTP-viðbót skjáborðs.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -345,14 +347,14 @@
   "settings.appearance.presetDefault": "Sjálfgefið",
   "settings.appearance.presetHint": "Forstillir lag á dökkt/ljóst val þitt. Writer's Sepia aðlagast að fullu að báðum stillingum.",
   "settings.appearance.presetSepia": "Sepia rithöfundar",
-  "settings.appearance.writingSurface": "Writing surface",
-  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
-  "settings.appearance.writingSurfaceTextured": "Textured",
-  "settings.appearance.writingSurfacePlain": "Plain",
   "settings.appearance.primaryColor": "Aðal litur",
   "settings.appearance.secondaryColor": "Secondary Litur",
   "settings.appearance.theme": "Umsóknarþema",
   "settings.appearance.title": "Útlit",
+  "settings.appearance.writingSurface": "Writing surface",
+  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
+  "settings.appearance.writingSurfacePlain": "Plain",
+  "settings.appearance.writingSurfaceTextured": "Textured",
   "settings.categories.about": "Um",
   "settings.categories.accessibility": "Aðgengi",
   "settings.categories.advancedAi": "Háþróuð gervigreind",
@@ -689,13 +691,13 @@
   "settings.privacy.encryptionSetAction": "Stilltu lykilorð",
   "settings.privacy.encryptionSetButton": "Virkja og dulkóða",
   "settings.privacy.encryptionSetup": "Stilltu lykilorð til að dulkóða öll handrit, skyndimyndir og stillingar sem vistaðar eru á þessu tæki.",
+  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.encryptionTooManyAttempts": "Of margar misheppnaðar tilraunir. Vinsamlegast bíddu í {{seconds}} sekúndur.",
   "settings.privacy.encryptionTooShort": "Aðgangsorð verður að vera að minnsta kosti 8 stafir",
   "settings.privacy.encryptionUnlockAction": "Opnaðu geymslu",
   "settings.privacy.encryptionUnlockButton": "Opnaðu",
   "settings.privacy.encryptionWarning": "Ef þú gleymir lykilorðinu þínu er ekki hægt að endurheimta gögnin þín.",
   "settings.privacy.encryptionWrongPassphrase": "Rangt lykilorð — afkóðun mistókst",
-  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.euDataResidency": "EU Gagnavistarleyfi",
   "settings.privacy.euDataResidencyHint": "Geymdu öll gögn innan netþjóna Evrópusambandsins þar sem við á.",
   "settings.privacy.localStorageOnly": "Aðeins staðbundin geymsla",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "Sækir {{model}} líkan ({{size}} MB). Þetta gerir staðbundna raddvinnslu kleift án skýjaleiðar.",
   "voice.modelDownload.error": "Niðurhal mistókst: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% lokið",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Reyndu aftur",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Raddlíkan niðurhal"
 }
diff --git a/locales/it/settings.json b/locales/it/settings.json
index 15974575..a730091f 100644
--- a/locales/it/settings.json
+++ b/locales/it/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "Download annullato",
   "settings.ai.localAi.downloadEta": "circa {{seconds}} s rimanenti",
   "settings.ai.localAi.downloadFailed": "Download non riuscito. Riprova.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB di {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Download del modello IA in corso…",
   "settings.ai.localAi.downloadedLabel": "scaricato",
   "settings.ai.localAi.downloadingButton": "Download in corso…",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "Download del modello {{model}} ({{size}} MB). Questo abilita l'elaborazione vocale locale senza instradamento cloud.",
   "voice.modelDownload.error": "Download fallito: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% completato",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB di {{total}} MB",
   "voice.modelDownload.retry": "Riprova",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Download modello vocale"
 }
diff --git a/locales/ja/settings.json b/locales/ja/settings.json
index 3f7f6109..55339c03 100644
--- a/locales/ja/settings.json
+++ b/locales/ja/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "ダウンロードがキャンセルされました",
   "settings.ai.localAi.downloadEta": "約。残り {{seconds}}秒",
   "settings.ai.localAi.downloadFailed": "ダウンロードに失敗しました。もう一度お試しください。",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "AI モデルをダウンロードしています…",
   "settings.ai.localAi.downloadedLabel": "ダウンロードされた",
   "settings.ai.localAi.downloadingButton": "ダウンロード中…",
@@ -198,6 +200,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "利用不可",
   "settings.ai.localBackendPreset": "ローカル バックエンド プリセット",
+  "settings.ai.localDiagnostic.browserFetch": "ブラウザー取得",
+  "settings.ai.localDiagnostic.endpoint": "エンドポイント",
+  "settings.ai.localDiagnostic.models": "モデル",
+  "settings.ai.localDiagnostic.tauriHttp": "Tauri ネイティブ HTTP",
+  "settings.ai.localDiagnostic.title": "接続診断",
+  "settings.ai.localDiagnostic.transport": "通信方式",
   "settings.ai.ollamaBrowserNote": "Ollama はデスクトップ アプリでのみ利用できます。ブラウザのコンテンツ セキュリティ ポリシーは、ローカルホストへの直接接続をブロックします。",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -229,15 +237,9 @@
   "settings.ai.providerStatusConnected": "接続済み",
   "settings.ai.providerStatusDisconnected": "利用不可",
   "settings.ai.providerStatusLabel": "接続ステータス",
-  "settings.ai.providerStatusReady": "準備ができて",
   "settings.ai.providerStatusNotTested": "未テスト",
+  "settings.ai.providerStatusReady": "準備ができて",
   "settings.ai.providerStatusTesting": "テスト中…",
-  "settings.ai.localDiagnostic.title": "接続診断",
-  "settings.ai.localDiagnostic.endpoint": "エンドポイント",
-  "settings.ai.localDiagnostic.transport": "通信方式",
-  "settings.ai.localDiagnostic.models": "モデル",
-  "settings.ai.localDiagnostic.tauriHttp": "Tauri ネイティブ HTTP",
-  "settings.ai.localDiagnostic.browserFetch": "ブラウザー取得",
   "settings.ai.providerStatusUnavailableBrowser": "ブラウザでは利用できません",
   "settings.ai.providerTitle": "AIプロバイダー",
   "settings.ai.providerTransformers": "トランスフォーマー.js",
@@ -260,8 +262,8 @@
   "settings.ai.testError.desktopRequired": "Ollama およびローカルの OpenAI 互換サーバーはデスクトップアプリでのみ利用できます。ブラウザは Web ページから localhost への直接接続をブロックします（CORS および Private Network Access）。",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "{{provider}} の API キーが設定されていません",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "このブラウザでは WebGPU が利用できません — WebLLM には WebGPU が必要です（Chrome/Edge を試すかフラグを有効にしてください）。",
   "settings.ai.testError.pluginUnavailable": "ローカルサーバーへの接続が利用できません: デスクトップ HTTP プラグインの読み込みに失敗しました。",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -345,14 +347,14 @@
   "settings.appearance.presetDefault": "デフォルト",
   "settings.appearance.presetHint": "プリセットはライト/ダークモードと組み合わさります。作家のセピアは両方のモードに完全対応しています。",
   "settings.appearance.presetSepia": "作家のセピア",
-  "settings.appearance.writingSurface": "執筆背景",
-  "settings.appearance.writingSurfaceHint": "長文執筆に最も心地よい背景スタイルを選んでください。",
-  "settings.appearance.writingSurfaceTextured": "テクスチャあり",
-  "settings.appearance.writingSurfacePlain": "無地",
   "settings.appearance.primaryColor": "原色",
   "settings.appearance.secondaryColor": "二次色",
   "settings.appearance.theme": "アプリケーションのテーマ",
   "settings.appearance.title": "外観",
+  "settings.appearance.writingSurface": "執筆背景",
+  "settings.appearance.writingSurfaceHint": "長文執筆に最も心地よい背景スタイルを選んでください。",
+  "settings.appearance.writingSurfacePlain": "無地",
+  "settings.appearance.writingSurfaceTextured": "テクスチャあり",
   "settings.categories.about": "について",
   "settings.categories.accessibility": "アクセシビリティ",
   "settings.categories.advancedAi": "高度なAI",
@@ -689,13 +691,13 @@
   "settings.privacy.encryptionSetAction": "パスフレーズを設定する",
   "settings.privacy.encryptionSetButton": "有効化と暗号化",
   "settings.privacy.encryptionSetup": "このデバイスに保存されているすべての原稿、スナップショット、設定を暗号化するためのパスフレーズを設定します。",
+  "settings.privacy.encryptionSetupFailed": "暗号化の設定を保存できませんでした — このブラウザでローカルストレージが許可されているか確認してから、もう一度お試しください",
   "settings.privacy.encryptionTooManyAttempts": "失敗した試行が多すぎます。 {{seconds}}秒待ってください。",
   "settings.privacy.encryptionTooShort": "パスフレーズは 8 文字以上である必要があります",
   "settings.privacy.encryptionUnlockAction": "ストレージのロックを解除する",
   "settings.privacy.encryptionUnlockButton": "ロックを解除する",
   "settings.privacy.encryptionWarning": "パスフレーズを忘れた場合、データは復元できません。",
   "settings.privacy.encryptionWrongPassphrase": "パスフレーズが間違っています - 復号化に失敗しました",
-  "settings.privacy.encryptionSetupFailed": "暗号化の設定を保存できませんでした — このブラウザでローカルストレージが許可されているか確認してから、もう一度お試しください",
   "settings.privacy.euDataResidency": "EUデータレジデンシー",
   "settings.privacy.euDataResidencyHint": "可能な限り、すべてのデータをEUサーバーに保持します。",
   "settings.privacy.localStorageOnly": "ローカルストレージのみ",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "{{model}} モデルをダウンロードしています ({{size}} MB)。これにより、クラウド ルーティングを使用せずにローカル音声処理が可能になります。",
   "voice.modelDownload.error": "ダウンロードに失敗しました: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% 完了",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "リトライ",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "音声モデルのダウンロード"
 }
diff --git a/locales/ko/settings.json b/locales/ko/settings.json
index 84548a36..5ad0cc9c 100644
--- a/locales/ko/settings.json
+++ b/locales/ko/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "다운로드가 취소되었습니다.",
   "settings.ai.localAi.downloadEta": "대략. {{seconds}}초 남음",
   "settings.ai.localAi.downloadFailed": "다운로드에 실패했습니다. 다시 시도해 주세요.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "AI 모델 다운로드 중…",
   "settings.ai.localAi.downloadedLabel": "다운로드됨",
   "settings.ai.localAi.downloadingButton": "다운로드 중…",
@@ -198,6 +200,12 @@
   "settings.ai.localAi.webgpuLabel": "웹GPU",
   "settings.ai.localAi.webgpuUnavailable": "사용할 수 없음",
   "settings.ai.localBackendPreset": "로컬 백엔드 사전 설정",
+  "settings.ai.localDiagnostic.browserFetch": "브라우저 가져오기",
+  "settings.ai.localDiagnostic.endpoint": "엔드포인트",
+  "settings.ai.localDiagnostic.models": "모델",
+  "settings.ai.localDiagnostic.tauriHttp": "Tauri 네이티브 HTTP",
+  "settings.ai.localDiagnostic.title": "연결 진단",
+  "settings.ai.localDiagnostic.transport": "전송 방식",
   "settings.ai.ollamaBrowserNote": "Ollama는 데스크톱 앱에서만 사용할 수 있습니다. 브라우저의 콘텐츠 보안 정책은 localhost에 대한 직접 연결을 차단합니다.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -229,15 +237,9 @@
   "settings.ai.providerStatusConnected": "연결됨",
   "settings.ai.providerStatusDisconnected": "없는",
   "settings.ai.providerStatusLabel": "연결 상태",
-  "settings.ai.providerStatusReady": "준비가 된",
   "settings.ai.providerStatusNotTested": "테스트되지 않음",
+  "settings.ai.providerStatusReady": "준비가 된",
   "settings.ai.providerStatusTesting": "테스트 중…",
-  "settings.ai.localDiagnostic.title": "연결 진단",
-  "settings.ai.localDiagnostic.endpoint": "엔드포인트",
-  "settings.ai.localDiagnostic.transport": "전송 방식",
-  "settings.ai.localDiagnostic.models": "모델",
-  "settings.ai.localDiagnostic.tauriHttp": "Tauri 네이티브 HTTP",
-  "settings.ai.localDiagnostic.browserFetch": "브라우저 가져오기",
   "settings.ai.providerStatusUnavailableBrowser": "브라우저에서 사용할 수 없음",
   "settings.ai.providerTitle": "AI 제공업체",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -260,8 +262,8 @@
   "settings.ai.testError.desktopRequired": "Ollama 및 로컬 OpenAI 호환 서버는 데스크톱 앱에서만 사용할 수 있습니다. 브라우저는 웹 페이지에서 localhost로의 직접 연결을 차단합니다(CORS 및 Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "{{provider}} API 키가 설정되지 않았습니다",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "이 브라우저에서는 WebGPU를 사용할 수 없습니다 — WebLLM에는 WebGPU가 필요합니다(Chrome/Edge를 사용하거나 플래그를 활성화하세요).",
   "settings.ai.testError.pluginUnavailable": "로컬 서버 네트워킹을 사용할 수 없습니다: 데스크톱 HTTP 플러그인을 불러오지 못했습니다.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -345,14 +347,14 @@
   "settings.appearance.presetDefault": "기본",
   "settings.appearance.presetHint": "어둡거나 밝은 선택으로 레이어를 사전 설정합니다. 작가의 세피아는 두 모드 모두에 완전히 적응합니다.",
   "settings.appearance.presetSepia": "작가의 세피아",
-  "settings.appearance.writingSurface": "글쓰기 배경",
-  "settings.appearance.writingSurfaceHint": "장문 집필에 가장 편안한 배경 스타일을 선택하세요.",
-  "settings.appearance.writingSurfaceTextured": "질감 있음",
-  "settings.appearance.writingSurfacePlain": "단색",
   "settings.appearance.primaryColor": "원색",
   "settings.appearance.secondaryColor": "보조 색상",
   "settings.appearance.theme": "응용 테마",
   "settings.appearance.title": "모습",
+  "settings.appearance.writingSurface": "글쓰기 배경",
+  "settings.appearance.writingSurfaceHint": "장문 집필에 가장 편안한 배경 스타일을 선택하세요.",
+  "settings.appearance.writingSurfacePlain": "단색",
+  "settings.appearance.writingSurfaceTextured": "질감 있음",
   "settings.categories.about": "에 대한",
   "settings.categories.accessibility": "접근성",
   "settings.categories.advancedAi": "고급 AI",
@@ -689,13 +691,13 @@
   "settings.privacy.encryptionSetAction": "암호 설정",
   "settings.privacy.encryptionSetButton": "활성화 및 암호화",
   "settings.privacy.encryptionSetup": "이 장치에 저장된 모든 원고, 스냅샷 및 설정을 암호화하려면 암호를 설정하세요.",
+  "settings.privacy.encryptionSetupFailed": "암호화 설정을 저장할 수 없습니다 — 이 브라우저가 로컬 저장소를 허용하는지 확인한 후 다시 시도하세요",
   "settings.privacy.encryptionTooManyAttempts": "실패한 시도가 너무 많습니다. {{seconds}}초만 기다려주세요.",
   "settings.privacy.encryptionTooShort": "암호는 8자 이상이어야 합니다.",
   "settings.privacy.encryptionUnlockAction": "스토리지 잠금 해제",
   "settings.privacy.encryptionUnlockButton": "터놓다",
   "settings.privacy.encryptionWarning": "암호를 잊어버리면 데이터를 복구할 수 없습니다.",
   "settings.privacy.encryptionWrongPassphrase": "잘못된 암호 - 암호 해독 실패",
-  "settings.privacy.encryptionSetupFailed": "암호화 설정을 저장할 수 없습니다 — 이 브라우저가 로컬 저장소를 허용하는지 확인한 후 다시 시도하세요",
   "settings.privacy.euDataResidency": "EU 데이터 상주",
   "settings.privacy.euDataResidencyHint": "해당하는 경우 모든 데이터를 유럽 연합 서버 내에 보관하십시오.",
   "settings.privacy.localStorageOnly": "로컬 저장소만",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "{{model}} 모델 다운로드 중({{size}}MB) 이를 통해 클라우드 라우팅 없이 로컬 음성 처리가 가능해집니다.",
   "voice.modelDownload.error": "다운로드 실패: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% 완료",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "다시 해 보다",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "음성 모델 다운로드"
 }
diff --git a/locales/pt/settings.json b/locales/pt/settings.json
index bb2ae258..179c9291 100644
--- a/locales/pt/settings.json
+++ b/locales/pt/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "Download cancelado",
   "settings.ai.localAi.downloadEta": "aprox. {{seconds}} restantes",
   "settings.ai.localAi.downloadFailed": "A transferência falhou. Tente novamente.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Downloading IA model…",
   "settings.ai.localAi.downloadedLabel": "baixado",
   "settings.ai.localAi.downloadingButton": "A transferir…",
@@ -198,6 +200,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Indisponível",
   "settings.ai.localBackendPreset": "Predefinição de back-end local",
+  "settings.ai.localDiagnostic.browserFetch": "Solicitação do navegador",
+  "settings.ai.localDiagnostic.endpoint": "Ponto de extremidade",
+  "settings.ai.localDiagnostic.models": "Modelos",
+  "settings.ai.localDiagnostic.tauriHttp": "HTTP nativo do Tauri",
+  "settings.ai.localDiagnostic.title": "Diagnóstico de conexão",
+  "settings.ai.localDiagnostic.transport": "Transporte",
   "settings.ai.ollamaBrowserNote": "Ollama está disponível apenas no aplicativo de desktop. A Política de Segurança de Conteúdo do navegador bloqueia conexões diretas com o host local.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -229,15 +237,9 @@
   "settings.ai.providerStatusConnected": "Conectado",
   "settings.ai.providerStatusDisconnected": "Indisponível",
   "settings.ai.providerStatusLabel": "Status da conexão",
-  "settings.ai.providerStatusReady": "Preparar",
   "settings.ai.providerStatusNotTested": "Não testado",
+  "settings.ai.providerStatusReady": "Preparar",
   "settings.ai.providerStatusTesting": "Testando…",
-  "settings.ai.localDiagnostic.title": "Diagnóstico de conexão",
-  "settings.ai.localDiagnostic.endpoint": "Ponto de extremidade",
-  "settings.ai.localDiagnostic.transport": "Transporte",
-  "settings.ai.localDiagnostic.models": "Modelos",
-  "settings.ai.localDiagnostic.tauriHttp": "HTTP nativo do Tauri",
-  "settings.ai.localDiagnostic.browserFetch": "Solicitação do navegador",
   "settings.ai.providerStatusUnavailableBrowser": "Não disponível no navegador",
   "settings.ai.providerTitle": "IA Provider",
   "settings.ai.providerTransformers": "Transformadores.js",
@@ -260,8 +262,8 @@
   "settings.ai.testError.desktopRequired": "O Ollama e os servidores locais compatíveis com OpenAI só estão disponíveis na aplicação de ambiente de trabalho. Os navegadores bloqueiam ligações diretas de páginas web ao localhost (CORS e Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "O servidor local devolveu uma resposta de modelos inválida.",
-  "settings.ai.testError.noModels": "O servidor local está acessível, mas não expõe modelos.",
   "settings.ai.testError.noApiKey": "Nenhuma chave de API {{provider}} definida",
+  "settings.ai.testError.noModels": "O servidor local está acessível, mas não expõe modelos.",
   "settings.ai.testError.noWebgpu": "WebGPU indisponível neste navegador — o WebLLM necessita de WebGPU (experimente o Chrome/Edge ou ative as flags).",
   "settings.ai.testError.pluginUnavailable": "A rede do servidor local não está disponível: o plugin HTTP de ambiente de trabalho falhou ao carregar.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -345,14 +347,14 @@
   "settings.appearance.presetDefault": "Padrão",
   "settings.appearance.presetHint": "Os presets se combinam com seu modo claro/escuro. O Sépia do escritor adapta-se totalmente a ambos os modos.",
   "settings.appearance.presetSepia": "Sépia do escritor",
-  "settings.appearance.writingSurface": "Superfície de escrita",
-  "settings.appearance.writingSurfaceHint": "Escolha o estilo de fundo mais confortável para escrita longa.",
-  "settings.appearance.writingSurfaceTextured": "Texturizada",
-  "settings.appearance.writingSurfacePlain": "Lisa",
   "settings.appearance.primaryColor": "Cor Primária",
   "settings.appearance.secondaryColor": "Cor Secundária",
   "settings.appearance.theme": "Tema do aplicativo",
   "settings.appearance.title": "Aparência",
+  "settings.appearance.writingSurface": "Superfície de escrita",
+  "settings.appearance.writingSurfaceHint": "Escolha o estilo de fundo mais confortável para escrita longa.",
+  "settings.appearance.writingSurfacePlain": "Lisa",
+  "settings.appearance.writingSurfaceTextured": "Texturizada",
   "settings.categories.about": "Sobre",
   "settings.categories.accessibility": "Acessibilidade",
   "settings.categories.advancedAi": "Advanced IA",
@@ -689,13 +691,13 @@
   "settings.privacy.encryptionSetAction": "Definir senha",
   "settings.privacy.encryptionSetButton": "Habilitar e criptografar",
   "settings.privacy.encryptionSetup": "Defina uma senha para criptografar todos os manuscritos, instantâneos e configurações armazenados neste dispositivo.",
+  "settings.privacy.encryptionSetupFailed": "Não foi possível salvar a configuração de criptografia — verifique se este navegador permite armazenamento local e tente novamente",
   "settings.privacy.encryptionTooManyAttempts": "Muitas tentativas fracassadas. Aguarde {{seconds}} segundos.",
   "settings.privacy.encryptionTooShort": "A senha deve ter pelo menos 8 caracteres",
   "settings.privacy.encryptionUnlockAction": "Desbloquear armazenamento",
   "settings.privacy.encryptionUnlockButton": "Desbloquear",
   "settings.privacy.encryptionWarning": "Se você esquecer sua senha, seus dados não poderão ser recuperados.",
   "settings.privacy.encryptionWrongPassphrase": "Senha errada – falha na descriptografia",
-  "settings.privacy.encryptionSetupFailed": "Não foi possível salvar a configuração de criptografia — verifique se este navegador permite armazenamento local e tente novamente",
   "settings.privacy.euDataResidency": "Residência de Dados na UE",
   "settings.privacy.euDataResidencyHint": "Manter todos os dados em servidores da União Europeia quando possível.",
   "settings.privacy.localStorageOnly": "Somente armazenamento local",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "Fazendo download do modelo {{model}} ({{size}} MB). Isso permite o processamento de voz local sem roteamento na nuvem.",
   "voice.modelDownload.error": "Falha no download: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% concluído",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Tentar novamente",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Download do modelo de voz"
 }
diff --git a/locales/ru/settings.json b/locales/ru/settings.json
index 39201c23..6862669c 100644
--- a/locales/ru/settings.json
+++ b/locales/ru/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "Загрузка отменена",
   "settings.ai.localAi.downloadEta": "ок. осталось {{seconds}}с",
   "settings.ai.localAi.downloadFailed": "Загрузка не удалась. Пожалуйста, попробуйте еще раз.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Загрузка модели ИИ…",
   "settings.ai.localAi.downloadedLabel": "скачал",
   "settings.ai.localAi.downloadingButton": "Загрузка…",
@@ -198,6 +200,12 @@
   "settings.ai.localAi.webgpuLabel": "Веб-графический процессор",
   "settings.ai.localAi.webgpuUnavailable": "Нет в наличии",
   "settings.ai.localBackendPreset": "Локальная настройка серверной части",
+  "settings.ai.localDiagnostic.browserFetch": "Запрос браузера",
+  "settings.ai.localDiagnostic.endpoint": "Конечная точка",
+  "settings.ai.localDiagnostic.models": "Модели",
+  "settings.ai.localDiagnostic.tauriHttp": "Нативный HTTP Tauri",
+  "settings.ai.localDiagnostic.title": "Диагностика подключения",
+  "settings.ai.localDiagnostic.transport": "Транспорт",
   "settings.ai.ollamaBrowserNote": "Оллама доступен только в настольном приложении. Политика безопасности контента браузера блокирует прямые подключения к локальному хосту.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -229,15 +237,9 @@
   "settings.ai.providerStatusConnected": "Подключено",
   "settings.ai.providerStatusDisconnected": "Недоступно",
   "settings.ai.providerStatusLabel": "Статус соединения",
-  "settings.ai.providerStatusReady": "Готовый",
   "settings.ai.providerStatusNotTested": "Не проверено",
+  "settings.ai.providerStatusReady": "Готовый",
   "settings.ai.providerStatusTesting": "Проверка…",
-  "settings.ai.localDiagnostic.title": "Диагностика подключения",
-  "settings.ai.localDiagnostic.endpoint": "Конечная точка",
-  "settings.ai.localDiagnostic.transport": "Транспорт",
-  "settings.ai.localDiagnostic.models": "Модели",
-  "settings.ai.localDiagnostic.tauriHttp": "Нативный HTTP Tauri",
-  "settings.ai.localDiagnostic.browserFetch": "Запрос браузера",
   "settings.ai.providerStatusUnavailableBrowser": "Недоступно в браузере",
   "settings.ai.providerTitle": "Поставщик ИИ",
   "settings.ai.providerTransformers": "Трансформеры.js",
@@ -260,8 +262,8 @@
   "settings.ai.testError.desktopRequired": "Ollama и локальные серверы, совместимые с OpenAI, доступны только в настольном приложении. Браузеры блокируют прямые подключения с веб-страниц к localhost (CORS и Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "Не задан ключ API {{provider}}",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU недоступен в этом браузере — WebLLM требует WebGPU (попробуйте Chrome/Edge или включите флаги).",
   "settings.ai.testError.pluginUnavailable": "Сеть локального сервера недоступна: не удалось загрузить настольный HTTP-плагин.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -345,14 +347,14 @@
   "settings.appearance.presetDefault": "По умолчанию",
   "settings.appearance.presetHint": "Слой предустановок по вашему выбору темный/светлый. Writer's Sepia полностью адаптируется к обоим режимам.",
   "settings.appearance.presetSepia": "Писательская сепия",
-  "settings.appearance.writingSurface": "Фон для письма",
-  "settings.appearance.writingSurfaceHint": "Выберите стиль фона, который наиболее комфортен для долгой работы с текстом.",
-  "settings.appearance.writingSurfaceTextured": "Текстурный",
-  "settings.appearance.writingSurfacePlain": "Однотонный",
   "settings.appearance.primaryColor": "Основной цвет",
   "settings.appearance.secondaryColor": "Вторичный цвет",
   "settings.appearance.theme": "Тема приложения",
   "settings.appearance.title": "Появление",
+  "settings.appearance.writingSurface": "Фон для письма",
+  "settings.appearance.writingSurfaceHint": "Выберите стиль фона, который наиболее комфортен для долгой работы с текстом.",
+  "settings.appearance.writingSurfacePlain": "Однотонный",
+  "settings.appearance.writingSurfaceTextured": "Текстурный",
   "settings.categories.about": "О",
   "settings.categories.accessibility": "Доступность",
   "settings.categories.advancedAi": "Продвинутый ИИ",
@@ -689,13 +691,13 @@
   "settings.privacy.encryptionSetAction": "Установить парольную фразу",
   "settings.privacy.encryptionSetButton": "Включить и зашифровать",
   "settings.privacy.encryptionSetup": "Установите парольную фразу для шифрования всех рукописей, снимков и настроек, хранящихся на этом устройстве.",
+  "settings.privacy.encryptionSetupFailed": "Не удалось сохранить настройку шифрования — убедитесь, что этот браузер разрешает локальное хранилище, и повторите попытку",
   "settings.privacy.encryptionTooManyAttempts": "Слишком много неудачных попыток. Пожалуйста, подождите {{seconds}} секунд.",
   "settings.privacy.encryptionTooShort": "Парольная фраза должна содержать не менее 8 символов.",
   "settings.privacy.encryptionUnlockAction": "Разблокировать хранилище",
   "settings.privacy.encryptionUnlockButton": "Разблокировать",
   "settings.privacy.encryptionWarning": "Если вы забудете парольную фразу, ваши данные не смогут быть восстановлены.",
   "settings.privacy.encryptionWrongPassphrase": "Неверная парольная фраза — расшифровка не удалась.",
-  "settings.privacy.encryptionSetupFailed": "Не удалось сохранить настройку шифрования — убедитесь, что этот браузер разрешает локальное хранилище, и повторите попытку",
   "settings.privacy.euDataResidency": "Резиденция данных ЕС",
   "settings.privacy.euDataResidencyHint": "Храните все данные на серверах Европейского Союза, где это применимо.",
   "settings.privacy.localStorageOnly": "Только локальное хранилище",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "Скачиваем модель {{model}} ({{size}} МБ). Это обеспечивает локальную обработку голоса без облачной маршрутизации.",
   "voice.modelDownload.error": "Загрузка не удалась: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% выполнено",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Повторить попытку",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Скачать модель голоса"
 }
diff --git a/locales/sv/settings.json b/locales/sv/settings.json
index e43e0b7e..da65676d 100644
--- a/locales/sv/settings.json
+++ b/locales/sv/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "Nedladdningen avbröts",
   "settings.ai.localAi.downloadEta": "ca. {{seconds}}s kvar",
   "settings.ai.localAi.downloadFailed": "Nedladdningen misslyckades. Försök igen.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Laddar ner AI-modell...",
   "settings.ai.localAi.downloadedLabel": "nedladdade",
   "settings.ai.localAi.downloadingButton": "Laddar ned...",
@@ -198,6 +200,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Ej tillgängligt",
   "settings.ai.localBackendPreset": "Lokal förinställning av backend",
+  "settings.ai.localDiagnostic.browserFetch": "Webbläsarhämtning",
+  "settings.ai.localDiagnostic.endpoint": "Slutpunkt",
+  "settings.ai.localDiagnostic.models": "Modeller",
+  "settings.ai.localDiagnostic.tauriHttp": "Tauris inbyggda HTTP",
+  "settings.ai.localDiagnostic.title": "Anslutningsdiagnostik",
+  "settings.ai.localDiagnostic.transport": "Transport",
   "settings.ai.ollamaBrowserNote": "Ollama är endast tillgänglig i skrivbordsappen. Webbläsarens innehållssäkerhetspolicy blockerar direktanslutningar till localhost.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -229,15 +237,9 @@
   "settings.ai.providerStatusConnected": "Ansluten",
   "settings.ai.providerStatusDisconnected": "Inte tillgänglig",
   "settings.ai.providerStatusLabel": "Anslutningsstatus",
-  "settings.ai.providerStatusReady": "Redo",
   "settings.ai.providerStatusNotTested": "Inte testad",
+  "settings.ai.providerStatusReady": "Redo",
   "settings.ai.providerStatusTesting": "Testar…",
-  "settings.ai.localDiagnostic.title": "Anslutningsdiagnostik",
-  "settings.ai.localDiagnostic.endpoint": "Slutpunkt",
-  "settings.ai.localDiagnostic.transport": "Transport",
-  "settings.ai.localDiagnostic.models": "Modeller",
-  "settings.ai.localDiagnostic.tauriHttp": "Tauris inbyggda HTTP",
-  "settings.ai.localDiagnostic.browserFetch": "Webbläsarhämtning",
   "settings.ai.providerStatusUnavailableBrowser": "Inte tillgängligt i webbläsaren",
   "settings.ai.providerTitle": "AI-leverantör",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -260,8 +262,8 @@
   "settings.ai.testError.desktopRequired": "Ollama och lokala OpenAI-kompatibla servrar är endast tillgängliga i skrivbordsappen. Webbläsare blockerar direkta anslutningar från webbsidor till localhost (CORS och Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "Ingen {{provider}}-API-nyckel angiven",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU är inte tillgängligt i den här webbläsaren — WebLLM behöver WebGPU (prova Chrome/Edge eller aktivera flaggor).",
   "settings.ai.testError.pluginUnavailable": "Lokal servernätverksanslutning ej tillgänglig: skrivbordets HTTP-plugin misslyckades med att läsas in.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -345,14 +347,14 @@
   "settings.appearance.presetDefault": "Standard",
   "settings.appearance.presetHint": "Förinställer lager på ditt mörka/ljusa val. Writer's Sepia anpassar sig helt till båda lägena.",
   "settings.appearance.presetSepia": "Författarens Sepia",
-  "settings.appearance.writingSurface": "Writing surface",
-  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
-  "settings.appearance.writingSurfaceTextured": "Textured",
-  "settings.appearance.writingSurfacePlain": "Plain",
   "settings.appearance.primaryColor": "Grundfärg",
   "settings.appearance.secondaryColor": "Sekundär färg",
   "settings.appearance.theme": "Programtema",
   "settings.appearance.title": "Utseende",
+  "settings.appearance.writingSurface": "Writing surface",
+  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
+  "settings.appearance.writingSurfacePlain": "Plain",
+  "settings.appearance.writingSurfaceTextured": "Textured",
   "settings.categories.about": "Om",
   "settings.categories.accessibility": "Tillgänglighet",
   "settings.categories.advancedAi": "Avancerad AI",
@@ -689,13 +691,13 @@
   "settings.privacy.encryptionSetAction": "Ställ in lösenordsfras",
   "settings.privacy.encryptionSetButton": "Aktivera och kryptera",
   "settings.privacy.encryptionSetup": "Ställ in en lösenordsfras för att kryptera alla manuskript, ögonblicksbilder och inställningar som lagras på den här enheten.",
+  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.encryptionTooManyAttempts": "För många misslyckade försök. Vänta {{seconds}} sekunder.",
   "settings.privacy.encryptionTooShort": "Lösenfrasen måste bestå av minst 8 tecken",
   "settings.privacy.encryptionUnlockAction": "Lås upp lagring",
   "settings.privacy.encryptionUnlockButton": "Låsa upp",
   "settings.privacy.encryptionWarning": "Om du glömmer din lösenfras kan din data inte återställas.",
   "settings.privacy.encryptionWrongPassphrase": "Fel lösenfras — dekrypteringen misslyckades",
-  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.euDataResidency": "EU Data Residency",
   "settings.privacy.euDataResidencyHint": "Behåll all data på Europeiska unionens servrar där det är tillämpligt.",
   "settings.privacy.localStorageOnly": "Endast lokal lagring",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "Laddar ner {{model}}-modellen ({{size}} MB). Detta möjliggör lokal röstbehandling utan molnrouting.",
   "voice.modelDownload.error": "Nedladdning misslyckades: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% klar",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Försöka igen",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Ladda ner röstmodell"
 }
diff --git a/locales/zh/settings.json b/locales/zh/settings.json
index 265294cb..52d9500b 100644
--- a/locales/zh/settings.json
+++ b/locales/zh/settings.json
@@ -159,6 +159,8 @@
   "settings.ai.localAi.downloadCancelled": "下载已取消",
   "settings.ai.localAi.downloadEta": "约。还剩 {{seconds}} 秒",
   "settings.ai.localAi.downloadFailed": "下载失败。请重试。",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "正在下载 AI 模型...",
   "settings.ai.localAi.downloadedLabel": "下载的",
   "settings.ai.localAi.downloadingButton": "下载中…",
@@ -198,6 +200,12 @@
   "settings.ai.localAi.webgpuLabel": "网络GPU",
   "settings.ai.localAi.webgpuUnavailable": "不可用",
   "settings.ai.localBackendPreset": "本地后端预设",
+  "settings.ai.localDiagnostic.browserFetch": "浏览器请求",
+  "settings.ai.localDiagnostic.endpoint": "端点",
+  "settings.ai.localDiagnostic.models": "模型",
+  "settings.ai.localDiagnostic.tauriHttp": "Tauri 原生 HTTP",
+  "settings.ai.localDiagnostic.title": "连接诊断",
+  "settings.ai.localDiagnostic.transport": "传输方式",
   "settings.ai.ollamaBrowserNote": "Ollama 仅在桌面应用程序中可用。浏览器的内容安全策略阻止与本地主机的直接连接。",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -229,15 +237,9 @@
   "settings.ai.providerStatusConnected": "已连接",
   "settings.ai.providerStatusDisconnected": "不可用",
   "settings.ai.providerStatusLabel": "连接状态",
-  "settings.ai.providerStatusReady": "准备好",
   "settings.ai.providerStatusNotTested": "未测试",
+  "settings.ai.providerStatusReady": "准备好",
   "settings.ai.providerStatusTesting": "测试中…",
-  "settings.ai.localDiagnostic.title": "连接诊断",
-  "settings.ai.localDiagnostic.endpoint": "端点",
-  "settings.ai.localDiagnostic.transport": "传输方式",
-  "settings.ai.localDiagnostic.models": "模型",
-  "settings.ai.localDiagnostic.tauriHttp": "Tauri 原生 HTTP",
-  "settings.ai.localDiagnostic.browserFetch": "浏览器请求",
   "settings.ai.providerStatusUnavailableBrowser": "浏览器中不可用",
   "settings.ai.providerTitle": "人工智能提供商",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -260,8 +262,8 @@
   "settings.ai.testError.desktopRequired": "Ollama 和本地 OpenAI 兼容服务器仅在桌面应用中可用。浏览器会阻止网页直接连接到 localhost（CORS 和 Private Network Access）。",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "未设置 {{provider}} API 密钥",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "此浏览器不支持 WebGPU — WebLLM 需要 WebGPU（请尝试 Chrome/Edge 或启用相关标志）。",
   "settings.ai.testError.pluginUnavailable": "本地服务器网络不可用：桌面 HTTP 插件加载失败。",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -345,14 +347,14 @@
   "settings.appearance.presetDefault": "默认",
   "settings.appearance.presetHint": "预设与深色/浅色模式叠加。作家的棕褐色可完全适应两种模式。",
   "settings.appearance.presetSepia": "作家的棕褐色",
-  "settings.appearance.writingSurface": "写作背景",
-  "settings.appearance.writingSurfaceHint": "选择最适合长篇写作的背景样式。",
-  "settings.appearance.writingSurfaceTextured": "有纹理",
-  "settings.appearance.writingSurfacePlain": "纯色",
   "settings.appearance.primaryColor": "原色",
   "settings.appearance.secondaryColor": "次要颜色",
   "settings.appearance.theme": "应用主题",
   "settings.appearance.title": "外貌",
+  "settings.appearance.writingSurface": "写作背景",
+  "settings.appearance.writingSurfaceHint": "选择最适合长篇写作的背景样式。",
+  "settings.appearance.writingSurfacePlain": "纯色",
+  "settings.appearance.writingSurfaceTextured": "有纹理",
   "settings.categories.about": "关于",
   "settings.categories.accessibility": "无障碍",
   "settings.categories.advancedAi": "先进的人工智能",
@@ -689,13 +691,13 @@
   "settings.privacy.encryptionSetAction": "设置密码",
   "settings.privacy.encryptionSetButton": "启用和加密",
   "settings.privacy.encryptionSetup": "设置密码以加密存储在该设备上的所有手稿、快照和设置。",
+  "settings.privacy.encryptionSetupFailed": "无法保存加密设置 — 请检查此浏览器是否允许本地存储，然后重试",
   "settings.privacy.encryptionTooManyAttempts": "失败的尝试太多了。请等待{{seconds}}秒。",
   "settings.privacy.encryptionTooShort": "密码必须至少为 8 个字符",
   "settings.privacy.encryptionUnlockAction": "解锁存储",
   "settings.privacy.encryptionUnlockButton": "开锁",
   "settings.privacy.encryptionWarning": "如果您忘记密码，您的数据将无法恢复。",
   "settings.privacy.encryptionWrongPassphrase": "密码错误 — 解密失败",
-  "settings.privacy.encryptionSetupFailed": "无法保存加密设置 — 请检查此浏览器是否允许本地存储，然后重试",
   "settings.privacy.euDataResidency": "欧盟数据驻留",
   "settings.privacy.euDataResidencyHint": "在可能的情况下，将所有数据保存在欧盟服务器上。",
   "settings.privacy.localStorageOnly": "仅本地存储",
@@ -819,6 +821,8 @@
   "voice.modelDownload.description": "正在下载 {{model}} 模型 ({{size}} MB)。这使得本地语音处理无需云路由。",
   "voice.modelDownload.error": "下载失败：{{error}}",
   "voice.modelDownload.progress": "{{percent}}% 完成",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "重试",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "语音模型下载"
 }
diff --git a/packages/ai-core/src/index.ts b/packages/ai-core/src/index.ts
index cde65dcc..22bc8265 100644
--- a/packages/ai-core/src/index.ts
+++ b/packages/ai-core/src/index.ts
@@ -182,6 +182,21 @@ export const WEBLLM_SUPPORTED_MODELS = [
 
 export type WebLlmModelId = (typeof WEBLLM_SUPPORTED_MODELS)[number]['id'];
 
+// QNBS-v3 (#333 item 1): machine-readable mirror of the approximate sizes already embedded as free
+// text in WEBLLM_SUPPORTED_MODELS' labels above (e.g. "~0.4 GB") — the installed @mlc-ai/web-llm's
+// own InitProgressReport type exposes only a 0-1 progress fraction + a status string, no structured
+// byte counts, so this is the only source for a downloaded/total MB estimate in the download UI.
+// Keep in sync with the labels above when adding/changing a supported model.
+export const WEBLLM_MODEL_APPROX_MB: Record<WebLlmModelId, number> = {
+  'Qwen2.5-0.5B-Instruct-q4f16_1-MLC': 400,
+  'Llama-3.2-1B-Instruct-q4f16_1-MLC': 700,
+  'Llama-3.2-3B-Instruct-q4f16_1-MLC': 1800,
+  'Phi-4-mini-instruct-q4f16_1-MLC': 2300,
+  'gemma-3-1b-it-q4f16_1-MLC': 800,
+  'gemma-3-4b-it-q4f32_1-MLC': 4900,
+  'Llama-3.3-70B-Instruct-q3f16_1-MLC': 35000,
+};
+
 export interface WebLlmProgressReport {
   progress: number; // 0–1
   text: string;
diff --git a/public/locales/ar/bundle.json b/public/locales/ar/bundle.json
index 718b5f6e..7b5af9f8 100644
--- a/public/locales/ar/bundle.json
+++ b/public/locales/ar/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "أُلغي التنزيل",
   "settings.ai.localAi.downloadEta": "يتبقّى ~{{seconds}} ثانية",
   "settings.ai.localAi.downloadFailed": "Download failed. Please try again.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "جارٍ تنزيل نموذج الذكاء الاصطناعي…",
   "settings.ai.localAi.downloadedLabel": "مُنزَّل",
   "settings.ai.localAi.downloadingButton": "Downloading…",
@@ -1844,6 +1846,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Not available",
   "settings.ai.localBackendPreset": "إعداد مسبق للخلفية المحلية",
+  "settings.ai.localDiagnostic.browserFetch": "طلب المتصفح",
+  "settings.ai.localDiagnostic.endpoint": "نقطة النهاية",
+  "settings.ai.localDiagnostic.models": "النماذج",
+  "settings.ai.localDiagnostic.tauriHttp": "HTTP الأصلي من Tauri",
+  "settings.ai.localDiagnostic.title": "تشخيص الاتصال",
+  "settings.ai.localDiagnostic.transport": "النقل",
   "settings.ai.ollamaBrowserNote": "‏Ollama متاح فقط في تطبيق سطح المكتب. تمنع سياسة أمان المحتوى للمتصفّح الاتصالات المباشرة بـ localhost.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -1875,15 +1883,9 @@
   "settings.ai.providerStatusConnected": "متصل",
   "settings.ai.providerStatusDisconnected": "غير متاح",
   "settings.ai.providerStatusLabel": "حالة الاتصال",
-  "settings.ai.providerStatusReady": "جاهز",
   "settings.ai.providerStatusNotTested": "لم يتم الاختبار",
+  "settings.ai.providerStatusReady": "جاهز",
   "settings.ai.providerStatusTesting": "جارٍ الاختبار…",
-  "settings.ai.localDiagnostic.title": "تشخيص الاتصال",
-  "settings.ai.localDiagnostic.endpoint": "نقطة النهاية",
-  "settings.ai.localDiagnostic.transport": "النقل",
-  "settings.ai.localDiagnostic.models": "النماذج",
-  "settings.ai.localDiagnostic.tauriHttp": "HTTP الأصلي من Tauri",
-  "settings.ai.localDiagnostic.browserFetch": "طلب المتصفح",
   "settings.ai.providerStatusUnavailableBrowser": "غير متوفر في المتصفح",
   "settings.ai.providerTitle": "مزوّد الذكاء الاصطناعي",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -1906,8 +1908,8 @@
   "settings.ai.testError.desktopRequired": "تتوفر Ollama والخوادم المحلية المتوافقة مع OpenAI فقط في تطبيق سطح المكتب. تحظر المتصفحات الاتصالات المباشرة من صفحات الويب بـ localhost (CORS وPrivate Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "لم يتم تعيين مفتاح API لـ {{provider}}",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU غير متوفر في هذا المتصفح — يحتاج WebLLM إلى WebGPU (جرّب Chrome/Edge أو فعّل العلامات).",
   "settings.ai.testError.pluginUnavailable": "شبكة الخادم المحلي غير متوفرة: فشل تحميل مكوّن HTTP لسطح المكتب.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -1991,14 +1993,14 @@
   "settings.appearance.presetDefault": "افتراضي",
   "settings.appearance.presetHint": "تتراكب الإعدادات المسبقة مع وضع الإضاءة. سيبيا الكاتب يتكيف مع كلا الوضعين.",
   "settings.appearance.presetSepia": "سيبيا الكاتب",
-  "settings.appearance.writingSurface": "سطح الكتابة",
-  "settings.appearance.writingSurfaceHint": "اختر نمط الخلفية الأنسب للكتابة الطويلة.",
-  "settings.appearance.writingSurfaceTextured": "منسوج",
-  "settings.appearance.writingSurfacePlain": "سادة",
   "settings.appearance.primaryColor": "اللون الأساسي",
   "settings.appearance.secondaryColor": "اللون الثانوي",
   "settings.appearance.theme": "سمة التطبيق",
   "settings.appearance.title": "المظهر",
+  "settings.appearance.writingSurface": "سطح الكتابة",
+  "settings.appearance.writingSurfaceHint": "اختر نمط الخلفية الأنسب للكتابة الطويلة.",
+  "settings.appearance.writingSurfacePlain": "سادة",
+  "settings.appearance.writingSurfaceTextured": "منسوج",
   "settings.categories.about": "حول",
   "settings.categories.accessibility": "إمكانية الوصول",
   "settings.categories.advancedAi": "ذكاء اصطناعي متقدّم",
@@ -2335,13 +2337,13 @@
   "settings.privacy.encryptionSetAction": "تعيين عبارة المرور",
   "settings.privacy.encryptionSetButton": "تفعيل وتشفير",
   "settings.privacy.encryptionSetup": "عيّن عبارة مرور لتشفير جميع المخطوطات واللقطات والإعدادات المُخزَّنة على هذا الجهاز.",
+  "settings.privacy.encryptionSetupFailed": "تعذر حفظ إعداد التشفير — تحقق من أن هذا المتصفح يسمح بالتخزين المحلي، ثم حاول مرة أخرى",
   "settings.privacy.encryptionTooManyAttempts": "محاولات فاشلة كثيرة جدًا. يرجى الانتظار {{seconds}} ثانية.",
   "settings.privacy.encryptionTooShort": "يجب ألا تقل عبارة المرور عن 8 أحرف",
   "settings.privacy.encryptionUnlockAction": "فتح التخزين",
   "settings.privacy.encryptionUnlockButton": "فتح",
   "settings.privacy.encryptionWarning": "إذا نسيت عبارة المرور، فلا يمكن استرجاع بياناتك.",
   "settings.privacy.encryptionWrongPassphrase": "عبارة مرور خاطئة — فشل فك التشفير",
-  "settings.privacy.encryptionSetupFailed": "تعذر حفظ إعداد التشفير — تحقق من أن هذا المتصفح يسمح بالتخزين المحلي، ثم حاول مرة أخرى",
   "settings.privacy.euDataResidency": "إقامة البيانات في الاتحاد الأوروبي",
   "settings.privacy.euDataResidencyHint": "EU Data Residency — keep all data within EU servers where applicable.",
   "settings.privacy.localStorageOnly": "التخزين المحلي فقط",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "جاري تحميل نموذج {{model}} ({{size}} ميغابايت). يُمكّن ذلك المعالجة الصوتية المحلية دون إرسال إلى السحابة.",
   "voice.modelDownload.error": "فشل التحميل: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% مكتمل",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "إعادة المحاولة",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "تحميل نموذج الصوت",
   "sidebar.characterGraph": "مخطط الشخصيات",
   "sidebar.characterInterviews": "مقابلات الشخصيات",
diff --git a/public/locales/de/bundle.json b/public/locales/de/bundle.json
index 4768536c..0fb36277 100644
--- a/public/locales/de/bundle.json
+++ b/public/locales/de/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "Download abgebrochen",
   "settings.ai.localAi.downloadEta": "ca. {{seconds}} Sek. verbleibend",
   "settings.ai.localAi.downloadFailed": "Download fehlgeschlagen. Bitte versuche es erneut.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB von {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "KI-Modell wird heruntergeladen…",
   "settings.ai.localAi.downloadedLabel": "heruntergeladen",
   "settings.ai.localAi.downloadingButton": "Wird heruntergeladen…",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "Lade {{model}}-Modell ({{size}} MB) herunter. Dies ermöglicht lokale Sprachverarbeitung ohne Cloud-Routing.",
   "voice.modelDownload.error": "Download fehlgeschlagen: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% abgeschlossen",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB von {{total}} MB",
   "voice.modelDownload.retry": "Wiederholen",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Sprachmodell-Download",
   "sidebar.characterGraph": "Charaktergraph",
   "sidebar.characterInterviews": "Charakterinterviews",
diff --git a/public/locales/el/bundle.json b/public/locales/el/bundle.json
index 40b034c1..3aa160b3 100644
--- a/public/locales/el/bundle.json
+++ b/public/locales/el/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "Η λήψη ακυρώθηκε",
   "settings.ai.localAi.downloadEta": "περίπου. Απομένουν {{seconds}}s",
   "settings.ai.localAi.downloadFailed": "Η λήψη απέτυχε. Δοκιμάστε ξανά.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Λήψη μοντέλου AI…",
   "settings.ai.localAi.downloadedLabel": "κατεβάσει",
   "settings.ai.localAi.downloadingButton": "Λήψη…",
@@ -1844,6 +1846,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Μη διαθέσιμο",
   "settings.ai.localBackendPreset": "Τοπική προεπιλογή υποστήριξης",
+  "settings.ai.localDiagnostic.browserFetch": "Αίτημα προγράμματος περιήγησης",
+  "settings.ai.localDiagnostic.endpoint": "Τελικό σημείο",
+  "settings.ai.localDiagnostic.models": "Μοντέλα",
+  "settings.ai.localDiagnostic.tauriHttp": "Εγγενές HTTP Tauri",
+  "settings.ai.localDiagnostic.title": "Διαγνωστικά σύνδεσης",
+  "settings.ai.localDiagnostic.transport": "Μεταφορά",
   "settings.ai.ollamaBrowserNote": "Το Olama είναι διαθέσιμο μόνο στην εφαρμογή για υπολογιστές. Η Πολιτική Ασφάλειας Περιεχομένου του προγράμματος περιήγησης αποκλείει τις άμεσες συνδέσεις με τον localhost.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -1875,15 +1883,9 @@
   "settings.ai.providerStatusConnected": "Συνδεδεμένος",
   "settings.ai.providerStatusDisconnected": "Δυσεύρετος",
   "settings.ai.providerStatusLabel": "Κατάσταση σύνδεσης",
-  "settings.ai.providerStatusReady": "Ετοιμος",
   "settings.ai.providerStatusNotTested": "Δεν έχει δοκιμαστεί",
+  "settings.ai.providerStatusReady": "Ετοιμος",
   "settings.ai.providerStatusTesting": "Δοκιμή σε εξέλιξη…",
-  "settings.ai.localDiagnostic.title": "Διαγνωστικά σύνδεσης",
-  "settings.ai.localDiagnostic.endpoint": "Τελικό σημείο",
-  "settings.ai.localDiagnostic.transport": "Μεταφορά",
-  "settings.ai.localDiagnostic.models": "Μοντέλα",
-  "settings.ai.localDiagnostic.tauriHttp": "Εγγενές HTTP Tauri",
-  "settings.ai.localDiagnostic.browserFetch": "Αίτημα προγράμματος περιήγησης",
   "settings.ai.providerStatusUnavailableBrowser": "Μη διαθέσιμο στο πρόγραμμα περιήγησης",
   "settings.ai.providerTitle": "Πάροχος AI",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -1906,8 +1908,8 @@
   "settings.ai.testError.desktopRequired": "Το Ollama και οι τοπικοί διακομιστές συμβατοί με OpenAI είναι διαθέσιμοι μόνο στην εφαρμογή επιφάνειας εργασίας. Οι φυλλομετρητές αποκλείουν άμεσες συνδέσεις από ιστοσελίδες προς το localhost (CORS και Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "Δεν έχει οριστεί κλειδί API {{provider}}",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "Το WebGPU δεν είναι διαθέσιμο σε αυτό το πρόγραμμα περιήγησης — το WebLLM απαιτεί WebGPU (δοκιμάστε Chrome/Edge ή ενεργοποιήστε τις σημαίες).",
   "settings.ai.testError.pluginUnavailable": "Το δίκτυο τοπικού διακομιστή δεν είναι διαθέσιμο: η πρόσθετη λειτουργία HTTP επιφάνειας εργασίας απέτυχε να φορτωθεί.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -1991,14 +1993,14 @@
   "settings.appearance.presetDefault": "Αθέτηση",
   "settings.appearance.presetHint": "Οι προεπιλογές συνδυάζονται με την επιλογή λειτουργίας. Η Σέπια του συγγραφέα προσαρμόζεται πλήρως και στις δύο λειτουργίες.",
   "settings.appearance.presetSepia": "Σέπια του συγγραφέα",
-  "settings.appearance.writingSurface": "Writing surface",
-  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
-  "settings.appearance.writingSurfaceTextured": "Textured",
-  "settings.appearance.writingSurfacePlain": "Plain",
   "settings.appearance.primaryColor": "Πρωτεύον χρώμα",
   "settings.appearance.secondaryColor": "Δευτερεύον Χρώμα",
   "settings.appearance.theme": "Θέμα εφαρμογής",
   "settings.appearance.title": "Εμφάνιση",
+  "settings.appearance.writingSurface": "Writing surface",
+  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
+  "settings.appearance.writingSurfacePlain": "Plain",
+  "settings.appearance.writingSurfaceTextured": "Textured",
   "settings.categories.about": "Για",
   "settings.categories.accessibility": "Προσιτότητα",
   "settings.categories.advancedAi": "Προηγμένη τεχνητή νοημοσύνη",
@@ -2335,13 +2337,13 @@
   "settings.privacy.encryptionSetAction": "Ορίστε τη φράση πρόσβασης",
   "settings.privacy.encryptionSetButton": "Ενεργοποίηση & Κρυπτογράφηση",
   "settings.privacy.encryptionSetup": "Ορίστε μια φράση πρόσβασης για να κρυπτογραφήσετε όλα τα χειρόγραφα, τα στιγμιότυπα και τις ρυθμίσεις που είναι αποθηκευμένα σε αυτήν τη συσκευή.",
+  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.encryptionTooManyAttempts": "Πάρα πολλές αποτυχημένες προσπάθειες. Περιμένετε {{seconds}} δευτερόλεπτα.",
   "settings.privacy.encryptionTooShort": "Η φράση πρόσβασης πρέπει να αποτελείται από τουλάχιστον 8 χαρακτήρες",
   "settings.privacy.encryptionUnlockAction": "Ξεκλείδωμα αποθήκευσης",
   "settings.privacy.encryptionUnlockButton": "Ξεκλειδώνω",
   "settings.privacy.encryptionWarning": "Εάν ξεχάσετε τη φράση πρόσβασης, τα δεδομένα σας δεν μπορούν να ανακτηθούν.",
   "settings.privacy.encryptionWrongPassphrase": "Λανθασμένη φράση πρόσβασης — η αποκρυπτογράφηση απέτυχε",
-  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.euDataResidency": "Εντοπισμός δεδομένων ΕΕ",
   "settings.privacy.euDataResidencyHint": "Διατήρηση όλων των δεδομένων σε servers της Ευρωπαϊκής Ένωσης όπου είναι δυνατόν.",
   "settings.privacy.localStorageOnly": "Μόνο τοπικός χώρος αποθήκευσης",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "Λήψη μοντέλου {{model}} ({{size}} MB). Αυτό επιτρέπει την τοπική επεξεργασία φωνής χωρίς δρομολόγηση cloud.",
   "voice.modelDownload.error": "Η λήψη απέτυχε: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% ολοκληρώθηκε",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Δοκιμάζω πάλι",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Λήψη φωνητικού μοντέλου",
   "sidebar.characterGraph": "Γράφημα χαρακτήρων",
   "sidebar.characterInterviews": "Συνεντεύξεις χαρακτήρων",
diff --git a/public/locales/en/bundle.json b/public/locales/en/bundle.json
index e906f12a..84c51d1d 100644
--- a/public/locales/en/bundle.json
+++ b/public/locales/en/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "Download cancelled",
   "settings.ai.localAi.downloadEta": "approx. {{seconds}}s remaining",
   "settings.ai.localAi.downloadFailed": "Download failed. Please try again.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Downloading AI model…",
   "settings.ai.localAi.downloadedLabel": "downloaded",
   "settings.ai.localAi.downloadingButton": "Downloading…",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "Downloading {{model}} model ({{size}} MB). This enables local voice processing without cloud routing.",
   "voice.modelDownload.error": "Download failed: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% complete",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Retry",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Voice Model Download",
   "sidebar.mainAria": "Main sidebar",
   "sidebar.primaryNavAria": "Primary navigation",
diff --git a/public/locales/es/bundle.json b/public/locales/es/bundle.json
index 1772ec5a..f9f6ec85 100644
--- a/public/locales/es/bundle.json
+++ b/public/locales/es/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "Descarga cancelada",
   "settings.ai.localAi.downloadEta": "aprox. {{seconds}} s restantes",
   "settings.ai.localAi.downloadFailed": "La descarga falló. Inténtalo de nuevo.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB de {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Descargando modelo de IA…",
   "settings.ai.localAi.downloadedLabel": "descargado",
   "settings.ai.localAi.downloadingButton": "Descargando…",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "Descargando modelo {{model}} ({{size}} MB). Esto habilita el procesamiento de voz local sin enrutamiento en la nube.",
   "voice.modelDownload.error": "Descarga fallida: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% completado",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB de {{total}} MB",
   "voice.modelDownload.retry": "Reintentar",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Descarga de modelo de voz",
   "sidebar.characterGraph": "Grafo de personajes",
   "sidebar.characterInterviews": "Entrevistas de personajes",
diff --git a/public/locales/eu/bundle.json b/public/locales/eu/bundle.json
index b7f5f0a8..d11d19a4 100644
--- a/public/locales/eu/bundle.json
+++ b/public/locales/eu/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "Deskarga bertan behera utzi da",
   "settings.ai.localAi.downloadEta": "gutxi gorabehera. {{seconds}}s geratzen dira",
   "settings.ai.localAi.downloadFailed": "Deskargatu huts egin du. Mesedez, saiatu berriro.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "AI eredua deskargatzen…",
   "settings.ai.localAi.downloadedLabel": "deskargatu",
   "settings.ai.localAi.downloadingButton": "Deskargatzen…",
@@ -1844,6 +1846,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Ez dago eskuragarri",
   "settings.ai.localBackendPreset": "Backend lokala aurrez ezarri",
+  "settings.ai.localDiagnostic.browserFetch": "Arakatzailearen eskaera",
+  "settings.ai.localDiagnostic.endpoint": "Amaiera-puntua",
+  "settings.ai.localDiagnostic.models": "Ereduak",
+  "settings.ai.localDiagnostic.tauriHttp": "Tauriren jatorrizko HTTPa",
+  "settings.ai.localDiagnostic.title": "Konexio-diagnostikoa",
+  "settings.ai.localDiagnostic.transport": "Garraioa",
   "settings.ai.ollamaBrowserNote": "Ollama mahaigaineko aplikazioan bakarrik dago eskuragarri. Arakatzailearen Edukiaren Segurtasun-Politikak lokaleko ostalari konexio zuzenak blokeatzen ditu.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -1875,15 +1883,9 @@
   "settings.ai.providerStatusConnected": "Konektatuta",
   "settings.ai.providerStatusDisconnected": "Ez dago erabilgarri",
   "settings.ai.providerStatusLabel": "Konexioaren egoera",
-  "settings.ai.providerStatusReady": "Prest",
   "settings.ai.providerStatusNotTested": "Probatu gabe",
+  "settings.ai.providerStatusReady": "Prest",
   "settings.ai.providerStatusTesting": "Probatzen…",
-  "settings.ai.localDiagnostic.title": "Konexio-diagnostikoa",
-  "settings.ai.localDiagnostic.endpoint": "Amaiera-puntua",
-  "settings.ai.localDiagnostic.transport": "Garraioa",
-  "settings.ai.localDiagnostic.models": "Ereduak",
-  "settings.ai.localDiagnostic.tauriHttp": "Tauriren jatorrizko HTTPa",
-  "settings.ai.localDiagnostic.browserFetch": "Arakatzailearen eskaera",
   "settings.ai.providerStatusUnavailableBrowser": "Ez dago erabilgarri nabigatzailean",
   "settings.ai.providerTitle": "AI hornitzailea",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -1906,8 +1908,8 @@
   "settings.ai.testError.desktopRequired": "Ollama eta OpenAI-rekin bateragarriak diren zerbitzari lokalak mahaigaineko aplikazioan bakarrik daude erabilgarri. Nabigatzaileek web orrialdeetatik localhost-era zuzeneko konexioak blokeatzen dituzte (CORS eta Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "Ez da {{provider}} API gakorik ezarri",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU ez dago erabilgarri nabigatzaile honetan — WebLLM-ek WebGPU behar du (probatu Chrome/Edge edo gaitu banderak).",
   "settings.ai.testError.pluginUnavailable": "Zerbitzari lokalaren sarea ez dago erabilgarri: mahaigaineko HTTP pluginak huts egin du kargatzean.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -1991,14 +1993,14 @@
   "settings.appearance.presetDefault": "Lehenetsia",
   "settings.appearance.presetHint": "Aurrez ezarri geruza zure aukera ilun/argian. Idazlearen Sepia guztiz egokitzen da bi moduetara.",
   "settings.appearance.presetSepia": "Idazlearen Sepia",
-  "settings.appearance.writingSurface": "Idazteko gainazala",
-  "settings.appearance.writingSurfaceHint": "Aukeratu testu luzeak idazteko erosoen zaizun atzeko planoaren estiloa.",
-  "settings.appearance.writingSurfaceTextured": "Testuratua",
-  "settings.appearance.writingSurfacePlain": "Lisoa",
   "settings.appearance.primaryColor": "Lehen kolorea",
   "settings.appearance.secondaryColor": "Bigarren mailako kolorea",
   "settings.appearance.theme": "Aplikazioaren gaia",
   "settings.appearance.title": "Itxura",
+  "settings.appearance.writingSurface": "Idazteko gainazala",
+  "settings.appearance.writingSurfaceHint": "Aukeratu testu luzeak idazteko erosoen zaizun atzeko planoaren estiloa.",
+  "settings.appearance.writingSurfacePlain": "Lisoa",
+  "settings.appearance.writingSurfaceTextured": "Testuratua",
   "settings.categories.about": "Buruz",
   "settings.categories.accessibility": "Irisgarritasuna",
   "settings.categories.advancedAi": "AI aurreratua",
@@ -2335,13 +2337,13 @@
   "settings.privacy.encryptionSetAction": "Ezarri pasaesaldia",
   "settings.privacy.encryptionSetButton": "Gaitu eta enkriptatu",
   "settings.privacy.encryptionSetup": "Ezarri pasaesaldi bat gailu honetan gordetako eskuizkribu, argazki eta ezarpen guztiak enkriptatzeko.",
+  "settings.privacy.encryptionSetupFailed": "Ezin izan da enkriptazioaren konfigurazioa gorde — egiaztatu nabigatzaile honek biltegiratze lokala onartzen duela, eta saiatu berriro",
   "settings.privacy.encryptionTooManyAttempts": "Porrot egindako saiakera gehiegi. Mesedez, itxaron {{seconds}} segundo.",
   "settings.privacy.encryptionTooShort": "Pasaesaldiek 8 karaktere izan behar dituzte gutxienez",
   "settings.privacy.encryptionUnlockAction": "Desblokeatu biltegia",
   "settings.privacy.encryptionUnlockButton": "Desblokeatu",
   "settings.privacy.encryptionWarning": "Pasaesaldia ahazten baduzu, ezin izango dira zure datuak berreskuratu.",
   "settings.privacy.encryptionWrongPassphrase": "Pasaesaldi okerra — deszifratzeak huts egin du",
-  "settings.privacy.encryptionSetupFailed": "Ezin izan da enkriptazioaren konfigurazioa gorde — egiaztatu nabigatzaile honek biltegiratze lokala onartzen duela, eta saiatu berriro",
   "settings.privacy.euDataResidency": "EBko Datuen Egoitza",
   "settings.privacy.euDataResidencyHint": "Gorde datu guztiak Europar Batasuneko zerbitzarietan, hala badagokio.",
   "settings.privacy.localStorageOnly": "Tokiko biltegiratzea soilik",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "{{model}} eredua deskargatzen ({{size}} MB). Honek ahots lokala prozesatzea ahalbidetzen du hodeiko bideratzerik gabe.",
   "voice.modelDownload.error": "Deskargatu huts egin du: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% osatuta",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Saiatu berriro",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Ahots eredua deskargatu",
   "sidebar.characterGraph": "Pertsonaia-grafikoa",
   "sidebar.characterInterviews": "Pertsonaia-elkarrizketak",
diff --git a/public/locales/fa/bundle.json b/public/locales/fa/bundle.json
index e95911fd..e15bd4c9 100644
--- a/public/locales/fa/bundle.json
+++ b/public/locales/fa/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "دانلود لغو شد",
   "settings.ai.localAi.downloadEta": "تقریبا {{seconds}} باقی مانده است",
   "settings.ai.localAi.downloadFailed": "دانلود انجام نشد. لطفا دوباره امتحان کنید.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "در حال دانلود مدل هوش مصنوعی…",
   "settings.ai.localAi.downloadedLabel": "دانلود شده است",
   "settings.ai.localAi.downloadingButton": "در حال دانلود…",
@@ -1844,6 +1846,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "در دسترس نیست",
   "settings.ai.localBackendPreset": "از پیش تنظیم شده باطن محلی",
+  "settings.ai.localDiagnostic.browserFetch": "دریافت مرورگر",
+  "settings.ai.localDiagnostic.endpoint": "نقطه پایانی",
+  "settings.ai.localDiagnostic.models": "مدل‌ها",
+  "settings.ai.localDiagnostic.tauriHttp": "HTTP بومی Tauri",
+  "settings.ai.localDiagnostic.title": "عیب‌یابی اتصال",
+  "settings.ai.localDiagnostic.transport": "انتقال",
   "settings.ai.ollamaBrowserNote": "Ollama فقط در برنامه دسکتاپ در دسترس است. خط مشی امنیت محتوای مرورگر اتصال مستقیم به لوکال هاست را مسدود می کند.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -1875,15 +1883,9 @@
   "settings.ai.providerStatusConnected": "متصل شد",
   "settings.ai.providerStatusDisconnected": "در دسترس نیست",
   "settings.ai.providerStatusLabel": "وضعیت اتصال",
-  "settings.ai.providerStatusReady": "آماده است",
   "settings.ai.providerStatusNotTested": "آزمایش نشده",
+  "settings.ai.providerStatusReady": "آماده است",
   "settings.ai.providerStatusTesting": "در حال آزمایش…",
-  "settings.ai.localDiagnostic.title": "عیب‌یابی اتصال",
-  "settings.ai.localDiagnostic.endpoint": "نقطه پایانی",
-  "settings.ai.localDiagnostic.transport": "انتقال",
-  "settings.ai.localDiagnostic.models": "مدل‌ها",
-  "settings.ai.localDiagnostic.tauriHttp": "HTTP بومی Tauri",
-  "settings.ai.localDiagnostic.browserFetch": "دریافت مرورگر",
   "settings.ai.providerStatusUnavailableBrowser": "در مرورگر در دسترس نیست",
   "settings.ai.providerTitle": "ارائه دهنده هوش مصنوعی",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -1906,8 +1908,8 @@
   "settings.ai.testError.desktopRequired": "Ollama و سرورهای محلی سازگار با OpenAI فقط در برنامه دسکتاپ در دسترس هستند. مرورگرها اتصالات مستقیم از صفحات وب به localhost را مسدود می‌کنند (CORS و Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "کلید API {{provider}} تنظیم نشده است",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU در این مرورگر در دسترس نیست — WebLLM به WebGPU نیاز دارد (Chrome/Edge را امتحان کنید یا پرچم‌ها را فعال کنید).",
   "settings.ai.testError.pluginUnavailable": "شبکه‌سازی سرور محلی در دسترس نیست: بارگذاری افزونه HTTP دسکتاپ ناموفق بود.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -1991,14 +1993,14 @@
   "settings.appearance.presetDefault": "پیش فرض",
   "settings.appearance.presetHint": "لایه از پیش تنظیم شده در انتخاب تاریک/روشن شما. Sepia نویسنده به طور کامل با هر دو حالت سازگار است.",
   "settings.appearance.presetSepia": "سپیای نویسنده",
-  "settings.appearance.writingSurface": "سطح نوشتن",
-  "settings.appearance.writingSurfaceHint": "سبک پس‌زمینه‌ای را که برای نوشتن طولانی راحت‌تر است انتخاب کنید.",
-  "settings.appearance.writingSurfaceTextured": "بافت‌دار",
-  "settings.appearance.writingSurfacePlain": "ساده",
   "settings.appearance.primaryColor": "رنگ اولیه",
   "settings.appearance.secondaryColor": "رنگ ثانویه",
   "settings.appearance.theme": "تم برنامه",
   "settings.appearance.title": "ظاهر",
+  "settings.appearance.writingSurface": "سطح نوشتن",
+  "settings.appearance.writingSurfaceHint": "سبک پس‌زمینه‌ای را که برای نوشتن طولانی راحت‌تر است انتخاب کنید.",
+  "settings.appearance.writingSurfacePlain": "ساده",
+  "settings.appearance.writingSurfaceTextured": "بافت‌دار",
   "settings.categories.about": "درباره",
   "settings.categories.accessibility": "قابلیت دسترسی",
   "settings.categories.advancedAi": "هوش مصنوعی پیشرفته",
@@ -2335,13 +2337,13 @@
   "settings.privacy.encryptionSetAction": "عبارت عبور را تنظیم کنید",
   "settings.privacy.encryptionSetButton": "فعال و رمزگذاری",
   "settings.privacy.encryptionSetup": "یک عبارت عبور برای رمزگذاری همه دست نوشته ها، عکس های فوری و تنظیمات ذخیره شده در این دستگاه تنظیم کنید.",
+  "settings.privacy.encryptionSetupFailed": "راه‌اندازی رمزگذاری ذخیره نشد — بررسی کنید که این مرورگر اجازه ذخیره‌سازی محلی را می‌دهد، سپس دوباره امتحان کنید",
   "settings.privacy.encryptionTooManyAttempts": "تلاش های ناموفق بسیار زیاد. لطفا {{seconds}} ثانیه صبر کنید.",
   "settings.privacy.encryptionTooShort": "عبارت عبور باید حداقل 8 کاراکتر باشد",
   "settings.privacy.encryptionUnlockAction": "قفل فضای ذخیره سازی را باز کنید",
   "settings.privacy.encryptionUnlockButton": "باز کردن قفل",
   "settings.privacy.encryptionWarning": "اگر رمز عبور خود را فراموش کنید، اطلاعات شما قابل بازیابی نیست.",
   "settings.privacy.encryptionWrongPassphrase": "عبارت عبور اشتباه - رمزگشایی انجام نشد",
-  "settings.privacy.encryptionSetupFailed": "راه‌اندازی رمزگذاری ذخیره نشد — بررسی کنید که این مرورگر اجازه ذخیره‌سازی محلی را می‌دهد، سپس دوباره امتحان کنید",
   "settings.privacy.euDataResidency": "اقامت داده اتحادیه اروپا",
   "settings.privacy.euDataResidencyHint": "در صورت لزوم، تمام داده ها را در سرورهای اتحادیه اروپا نگه دارید.",
   "settings.privacy.localStorageOnly": "فقط فضای ذخیره‌سازی محلی",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "در حال دانلود مدل {{model}} ({{size}} مگابایت). این امکان پردازش صدای محلی را بدون مسیریابی ابری فراهم می کند.",
   "voice.modelDownload.error": "دانلود انجام نشد: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% کامل شده است",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "دوباره امتحان کنید",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "دانلود مدل صوتی",
   "sidebar.characterGraph": "نمودار شخصیت‌ها",
   "sidebar.characterInterviews": "مصاحبه با شخصیت‌ها",
diff --git a/public/locales/fi/bundle.json b/public/locales/fi/bundle.json
index 9a5dff75..b2889564 100644
--- a/public/locales/fi/bundle.json
+++ b/public/locales/fi/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "Lataus peruutettu",
   "settings.ai.localAi.downloadEta": "noin {{seconds}}s jäljellä",
   "settings.ai.localAi.downloadFailed": "Lataus epäonnistui. Yritä uudelleen.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Ladataan tekoälymallia…",
   "settings.ai.localAi.downloadedLabel": "ladattu",
   "settings.ai.localAi.downloadingButton": "Ladataan…",
@@ -1844,6 +1846,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Ei saatavilla",
   "settings.ai.localBackendPreset": "Paikallinen taustajärjestelmän esiasetus",
+  "settings.ai.localDiagnostic.browserFetch": "Selaimen pyyntö",
+  "settings.ai.localDiagnostic.endpoint": "Päätepiste",
+  "settings.ai.localDiagnostic.models": "Mallit",
+  "settings.ai.localDiagnostic.tauriHttp": "Taurin natiivi HTTP",
+  "settings.ai.localDiagnostic.title": "Yhteysdiagnostiikka",
+  "settings.ai.localDiagnostic.transport": "Siirto",
   "settings.ai.ollamaBrowserNote": "Ollama on saatavilla vain työpöytäsovelluksessa. Selaimen sisällön suojauskäytäntö estää suorat yhteydet localhostiin.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -1875,15 +1883,9 @@
   "settings.ai.providerStatusConnected": "Yhdistetty",
   "settings.ai.providerStatusDisconnected": "Ei saatavilla",
   "settings.ai.providerStatusLabel": "Yhteyden tila",
-  "settings.ai.providerStatusReady": "Valmis",
   "settings.ai.providerStatusNotTested": "Ei testattu",
+  "settings.ai.providerStatusReady": "Valmis",
   "settings.ai.providerStatusTesting": "Testataan…",
-  "settings.ai.localDiagnostic.title": "Yhteysdiagnostiikka",
-  "settings.ai.localDiagnostic.endpoint": "Päätepiste",
-  "settings.ai.localDiagnostic.transport": "Siirto",
-  "settings.ai.localDiagnostic.models": "Mallit",
-  "settings.ai.localDiagnostic.tauriHttp": "Taurin natiivi HTTP",
-  "settings.ai.localDiagnostic.browserFetch": "Selaimen pyyntö",
   "settings.ai.providerStatusUnavailableBrowser": "Ei saatavilla selaimessa",
   "settings.ai.providerTitle": "AI Provider",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -1906,8 +1908,8 @@
   "settings.ai.testError.desktopRequired": "Ollama ja paikalliset OpenAI-yhteensopivat palvelimet ovat käytettävissä vain työpöytäsovelluksessa. Selaimet estävät suorat yhteydet verkkosivuilta osoitteeseen localhost (CORS ja Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "{{provider}}-API-avainta ei ole asetettu",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU ei ole käytettävissä tässä selaimessa — WebLLM tarvitsee WebGPU:n (kokeile Chromea/Edgeä tai ota liput käyttöön).",
   "settings.ai.testError.pluginUnavailable": "Paikallisen palvelimen verkko ei ole käytettävissä: työpöydän HTTP-lisäosan lataus epäonnistui.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -1991,14 +1993,14 @@
   "settings.appearance.presetDefault": "Oletus",
   "settings.appearance.presetHint": "Esiasettaa kerroksen tumman/vaalean valinnan mukaan. Writer's Sepia mukautuu täysin molempiin tiloihin.",
   "settings.appearance.presetSepia": "Kirjailijan Seepia",
-  "settings.appearance.writingSurface": "Writing surface",
-  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
-  "settings.appearance.writingSurfaceTextured": "Textured",
-  "settings.appearance.writingSurfacePlain": "Plain",
   "settings.appearance.primaryColor": "Ensisijainen väri",
   "settings.appearance.secondaryColor": "Toissijainen väri",
   "settings.appearance.theme": "Sovelluksen teema",
   "settings.appearance.title": "Ulkonäkö",
+  "settings.appearance.writingSurface": "Writing surface",
+  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
+  "settings.appearance.writingSurfacePlain": "Plain",
+  "settings.appearance.writingSurfaceTextured": "Textured",
   "settings.categories.about": "Noin",
   "settings.categories.accessibility": "Esteettömyys",
   "settings.categories.advancedAi": "Kehittynyt AI",
@@ -2335,13 +2337,13 @@
   "settings.privacy.encryptionSetAction": "Aseta salasana",
   "settings.privacy.encryptionSetButton": "Ota käyttöön ja salaa",
   "settings.privacy.encryptionSetup": "Aseta tunnuslause, joka salaa kaikki tähän laitteeseen tallennetut käsikirjoitukset, tilannevedokset ja asetukset.",
+  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.encryptionTooManyAttempts": "Liian monta epäonnistunutta yritystä. Odota {{seconds}} sekuntia.",
   "settings.privacy.encryptionTooShort": "Tunnuslauseessa on oltava vähintään 8 merkkiä",
   "settings.privacy.encryptionUnlockAction": "Avaa tallennustila",
   "settings.privacy.encryptionUnlockButton": "Avata",
   "settings.privacy.encryptionWarning": "Jos unohdat salasanasi, tietojasi ei voida palauttaa.",
   "settings.privacy.encryptionWrongPassphrase": "Väärä tunnuslause – salauksen purku epäonnistui",
-  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.euDataResidency": "EU Data Residency",
   "settings.privacy.euDataResidencyHint": "Säilytä kaikki tiedot tarvittaessa Euroopan unionin palvelimilla.",
   "settings.privacy.localStorageOnly": "Vain paikallinen tallennustila",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "Ladataan mallia {{model}} ({{size}} Mt). Tämä mahdollistaa paikallisen äänenkäsittelyn ilman pilvireititystä.",
   "voice.modelDownload.error": "Lataus epäonnistui: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% valmiina",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Yritä uudelleen",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Äänimallin lataus",
   "sidebar.characterGraph": "Hahmokaavio",
   "sidebar.characterInterviews": "Hahmohaastattelut",
diff --git a/public/locales/fr/bundle.json b/public/locales/fr/bundle.json
index 1e77267c..82cdb4bd 100644
--- a/public/locales/fr/bundle.json
+++ b/public/locales/fr/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "Téléchargement annulé",
   "settings.ai.localAi.downloadEta": "env. {{seconds}} s restantes",
   "settings.ai.localAi.downloadFailed": "Échec du téléchargement. Veuillez réessayer.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} Mo sur {{total}} Mo",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} Mo/s",
   "settings.ai.localAi.downloadTitle": "Téléchargement du modèle IA…",
   "settings.ai.localAi.downloadedLabel": "téléchargé",
   "settings.ai.localAi.downloadingButton": "Téléchargement…",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "Téléchargement du modèle {{model}} ({{size}} Mo). Cela permet le traitement vocal local sans routage cloud.",
   "voice.modelDownload.error": "Échec du téléchargement : {{error}}",
   "voice.modelDownload.progress": "{{percent}}% terminé",
+  "voice.modelDownload.progressBytes": "{{loaded}} Mo sur {{total}} Mo",
   "voice.modelDownload.retry": "Réessayer",
+  "voice.modelDownload.speed": "{{speed}} Mo/s",
   "voice.modelDownload.title": "Téléchargement du modèle vocal",
   "sidebar.characterGraph": "Graphe des personnages",
   "sidebar.characterInterviews": "Entretiens de personnages",
diff --git a/public/locales/he/bundle.json b/public/locales/he/bundle.json
index 42f10599..5a38fbe9 100644
--- a/public/locales/he/bundle.json
+++ b/public/locales/he/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "ההורדה בוטלה",
   "settings.ai.localAi.downloadEta": "נותרו כ‑{{seconds}} שנ׳",
   "settings.ai.localAi.downloadFailed": "Download failed. Please try again.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "מוריד מודל AI…",
   "settings.ai.localAi.downloadedLabel": "הורד",
   "settings.ai.localAi.downloadingButton": "Downloading…",
@@ -1844,6 +1846,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Not available",
   "settings.ai.localBackendPreset": "הגדרה מוגדרת מראש של מנוע אחורי מקומי",
+  "settings.ai.localDiagnostic.browserFetch": "בקשת דפדפן",
+  "settings.ai.localDiagnostic.endpoint": "נקודת קצה",
+  "settings.ai.localDiagnostic.models": "מודלים",
+  "settings.ai.localDiagnostic.tauriHttp": "HTTP מקורי של Tauri",
+  "settings.ai.localDiagnostic.title": "אבחון חיבור",
+  "settings.ai.localDiagnostic.transport": "תעבורה",
   "settings.ai.ollamaBrowserNote": "‏Ollama זמין רק באפליקציית שולחן העבודה. מדיניות אבטחת התוכן של הדפדפן חוסמת חיבורים ישירים ל‑localhost.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -1875,15 +1883,9 @@
   "settings.ai.providerStatusConnected": "מחובר",
   "settings.ai.providerStatusDisconnected": "לא זמין",
   "settings.ai.providerStatusLabel": "סטטוס חיבור",
-  "settings.ai.providerStatusReady": "מוכן",
   "settings.ai.providerStatusNotTested": "לא נבדק",
+  "settings.ai.providerStatusReady": "מוכן",
   "settings.ai.providerStatusTesting": "בודק…",
-  "settings.ai.localDiagnostic.title": "אבחון חיבור",
-  "settings.ai.localDiagnostic.endpoint": "נקודת קצה",
-  "settings.ai.localDiagnostic.transport": "תעבורה",
-  "settings.ai.localDiagnostic.models": "מודלים",
-  "settings.ai.localDiagnostic.tauriHttp": "HTTP מקורי של Tauri",
-  "settings.ai.localDiagnostic.browserFetch": "בקשת דפדפן",
   "settings.ai.providerStatusUnavailableBrowser": "לא זמין בדפדפן",
   "settings.ai.providerTitle": "ספק AI",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -1906,8 +1908,8 @@
   "settings.ai.testError.desktopRequired": "Ollama ושרתים מקומיים תואמי OpenAI זמינים רק באפליקציית שולחן העבודה. דפדפנים חוסמים חיבורים ישירים מדפי אינטרנט אל localhost (CORS ו-Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "לא הוגדר מפתח API עבור {{provider}}",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU אינו זמין בדפדפן זה — WebLLM דורש WebGPU (נסה Chrome/Edge או הפעל דגלים).",
   "settings.ai.testError.pluginUnavailable": "רשת השרת המקומי אינה זמינה: תוסף ה-HTTP של שולחן העבודה נכשל בטעינה.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -1991,14 +1993,14 @@
   "settings.appearance.presetDefault": "ברירת מחדל",
   "settings.appearance.presetHint": "ההגדרות המוגדרות מראש מצטברות עם בחירת המצב. ספיה של הכותב מסתגל לשני המצבים.",
   "settings.appearance.presetSepia": "ספיה של הכותב",
-  "settings.appearance.writingSurface": "משטח כתיבה",
-  "settings.appearance.writingSurfaceHint": "בחרו את סגנון הרקע הנוח ביותר לכתיבה ארוכה.",
-  "settings.appearance.writingSurfaceTextured": "בעל מרקם",
-  "settings.appearance.writingSurfacePlain": "חלק",
   "settings.appearance.primaryColor": "צבע ראשי",
   "settings.appearance.secondaryColor": "צבע משני",
   "settings.appearance.theme": "ערכת נושא של האפליקציה",
   "settings.appearance.title": "מראה",
+  "settings.appearance.writingSurface": "משטח כתיבה",
+  "settings.appearance.writingSurfaceHint": "בחרו את סגנון הרקע הנוח ביותר לכתיבה ארוכה.",
+  "settings.appearance.writingSurfacePlain": "חלק",
+  "settings.appearance.writingSurfaceTextured": "בעל מרקם",
   "settings.categories.about": "אודות",
   "settings.categories.accessibility": "נגישות",
   "settings.categories.advancedAi": "AI מתקדם",
@@ -2335,13 +2337,13 @@
   "settings.privacy.encryptionSetAction": "הגדרת ביטוי סיסמה",
   "settings.privacy.encryptionSetButton": "הפעלה והצפנה",
   "settings.privacy.encryptionSetup": "הגדירו ביטוי סיסמה כדי להצפין את כל כתבי היד, תמונות המצב וההגדרות המאוחסנים במכשיר זה.",
+  "settings.privacy.encryptionSetupFailed": "לא ניתן היה לשמור את הגדרת ההצפנה — ודאו שדפדפן זה מאפשר אחסון מקומי ונסו שוב",
   "settings.privacy.encryptionTooManyAttempts": "יותר מדי ניסיונות כושלים. אנא המתינו {{seconds}} שניות.",
   "settings.privacy.encryptionTooShort": "ביטוי הסיסמה חייב להכיל לפחות 8 תווים",
   "settings.privacy.encryptionUnlockAction": "פתיחת אחסון",
   "settings.privacy.encryptionUnlockButton": "פתיחה",
   "settings.privacy.encryptionWarning": "אם תשכחו את ביטוי הסיסמה, לא ניתן יהיה לשחזר את הנתונים שלכם.",
   "settings.privacy.encryptionWrongPassphrase": "ביטוי סיסמה שגוי — הפענוח נכשל",
-  "settings.privacy.encryptionSetupFailed": "לא ניתן היה לשמור את הגדרת ההצפנה — ודאו שדפדפן זה מאפשר אחסון מקומי ונסו שוב",
   "settings.privacy.euDataResidency": "אחסון נתונים באיחוד האירופי",
   "settings.privacy.euDataResidencyHint": "EU Data Residency — keep all data within EU servers where applicable.",
   "settings.privacy.localStorageOnly": "אחסון מקומי בלבד",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "מוריד את מודל {{model}} ({{size}} מגהבייט). זה מאפשר עיבוד קול מקומי ללא ניתוב לענן.",
   "voice.modelDownload.error": "ההורדה נכשלה: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% הושלם",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "נסה שוב",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "הורדת מודל קול",
   "sidebar.characterGraph": "גרף דמויות",
   "sidebar.characterInterviews": "ראיונות דמויות",
diff --git a/public/locales/hu/bundle.json b/public/locales/hu/bundle.json
index ba63457a..09ed79d1 100644
--- a/public/locales/hu/bundle.json
+++ b/public/locales/hu/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "Letöltés megszakítva",
   "settings.ai.localAi.downloadEta": "kb. {{seconds}}s van hátra",
   "settings.ai.localAi.downloadFailed": "Letöltés sikertelen. Kérjük, próbálja újra.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "AI-modell letöltése…",
   "settings.ai.localAi.downloadedLabel": "letöltve",
   "settings.ai.localAi.downloadingButton": "Letöltés…",
@@ -1844,6 +1846,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Nem elérhető",
   "settings.ai.localBackendPreset": "Helyi háttér előre beállított",
+  "settings.ai.localDiagnostic.browserFetch": "Böngészőlekérés",
+  "settings.ai.localDiagnostic.endpoint": "Végpont",
+  "settings.ai.localDiagnostic.models": "Modellek",
+  "settings.ai.localDiagnostic.tauriHttp": "Tauri natív HTTP",
+  "settings.ai.localDiagnostic.title": "Kapcsolati diagnosztika",
+  "settings.ai.localDiagnostic.transport": "Átvitel",
   "settings.ai.ollamaBrowserNote": "Az Ollama csak az asztali alkalmazásban érhető el. A böngésző tartalombiztonsági házirendje blokkolja a közvetlen kapcsolatot a localhosttal.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -1875,15 +1883,9 @@
   "settings.ai.providerStatusConnected": "Csatlakozva",
   "settings.ai.providerStatusDisconnected": "Nem elérhető",
   "settings.ai.providerStatusLabel": "Kapcsolat állapota",
-  "settings.ai.providerStatusReady": "Kész",
   "settings.ai.providerStatusNotTested": "Nincs tesztelve",
+  "settings.ai.providerStatusReady": "Kész",
   "settings.ai.providerStatusTesting": "Tesztelés…",
-  "settings.ai.localDiagnostic.title": "Kapcsolati diagnosztika",
-  "settings.ai.localDiagnostic.endpoint": "Végpont",
-  "settings.ai.localDiagnostic.transport": "Átvitel",
-  "settings.ai.localDiagnostic.models": "Modellek",
-  "settings.ai.localDiagnostic.tauriHttp": "Tauri natív HTTP",
-  "settings.ai.localDiagnostic.browserFetch": "Böngészőlekérés",
   "settings.ai.providerStatusUnavailableBrowser": "Böngészőben nem érhető el",
   "settings.ai.providerTitle": "AI szolgáltató",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -1906,8 +1908,8 @@
   "settings.ai.testError.desktopRequired": "Az Ollama és a helyi OpenAI-kompatibilis kiszolgálók csak az asztali alkalmazásban érhetők el. A böngészők blokkolják a weboldalak közvetlen kapcsolatait a localhosttal (CORS és Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "Nincs beállítva {{provider}} API-kulcs",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "A WebGPU nem érhető el ebben a böngészőben — a WebLLM-nek WebGPU-ra van szüksége (próbáld a Chrome-ot/Edge-et, vagy engedélyezd a jelzőket).",
   "settings.ai.testError.pluginUnavailable": "A helyi kiszolgálóhálózat nem érhető el: az asztali HTTP bővítmény betöltése sikertelen.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -1991,14 +1993,14 @@
   "settings.appearance.presetDefault": "Alapértelmezett",
   "settings.appearance.presetHint": "Előre beállítja a réteget a sötét/világos választás szerint. A Writer's Sepia teljes mértékben alkalmazkodik mindkét módhoz.",
   "settings.appearance.presetSepia": "Írói szépia",
-  "settings.appearance.writingSurface": "Writing surface",
-  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
-  "settings.appearance.writingSurfaceTextured": "Textured",
-  "settings.appearance.writingSurfacePlain": "Plain",
   "settings.appearance.primaryColor": "Elsődleges szín",
   "settings.appearance.secondaryColor": "Másodlagos szín",
   "settings.appearance.theme": "Alkalmazási téma",
   "settings.appearance.title": "Megjelenés",
+  "settings.appearance.writingSurface": "Writing surface",
+  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
+  "settings.appearance.writingSurfacePlain": "Plain",
+  "settings.appearance.writingSurfaceTextured": "Textured",
   "settings.categories.about": "Körülbelül",
   "settings.categories.accessibility": "Megközelíthetőség",
   "settings.categories.advancedAi": "Fejlett AI",
@@ -2335,13 +2337,13 @@
   "settings.privacy.encryptionSetAction": "Állítsa be a jelszót",
   "settings.privacy.encryptionSetButton": "Engedélyezés és titkosítás",
   "settings.privacy.encryptionSetup": "Állítson be egy jelszót az eszközön tárolt összes kézirat, pillanatkép és beállítás titkosításához.",
+  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.encryptionTooManyAttempts": "Túl sok sikertelen próbálkozás. Kérjük, várjon {{seconds}} másodpercet.",
   "settings.privacy.encryptionTooShort": "A jelszónak legalább 8 karakterből kell állnia",
   "settings.privacy.encryptionUnlockAction": "Tárolás feloldása",
   "settings.privacy.encryptionUnlockButton": "Kinyit",
   "settings.privacy.encryptionWarning": "Ha elfelejti jelszavát, adatait nem lehet visszaállítani.",
   "settings.privacy.encryptionWrongPassphrase": "Hibás jelszó – a visszafejtés nem sikerült",
-  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.euDataResidency": "EU Data Residency",
   "settings.privacy.euDataResidencyHint": "Adott esetben minden adatot tároljon az Európai Unió szerverein.",
   "settings.privacy.localStorageOnly": "Csak helyi tárhely",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "{{model}} modell letöltése ({{size}} MB). Ez lehetővé teszi a helyi hangfeldolgozást felhőalapú útválasztás nélkül.",
   "voice.modelDownload.error": "Sikertelen letöltés: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% kész",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Próbálja újra",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Hangmodell letöltése",
   "sidebar.characterGraph": "Karaktergráf",
   "sidebar.characterInterviews": "Karakterinterjúk",
diff --git a/public/locales/is/bundle.json b/public/locales/is/bundle.json
index cfc496bd..faf3840c 100644
--- a/public/locales/is/bundle.json
+++ b/public/locales/is/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "Hætt við niðurhal",
   "settings.ai.localAi.downloadEta": "ca. {{seconds}}s eftir",
   "settings.ai.localAi.downloadFailed": "Niðurhal mistókst. Vinsamlegast reyndu aftur.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Sækir gervigreind líkan…",
   "settings.ai.localAi.downloadedLabel": "niðurhalað",
   "settings.ai.localAi.downloadingButton": "Sækir…",
@@ -1844,6 +1846,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Ekki í boði",
   "settings.ai.localBackendPreset": "Forstilla staðbundin bakenda",
+  "settings.ai.localDiagnostic.browserFetch": "Vafrafyrirspurn",
+  "settings.ai.localDiagnostic.endpoint": "Endapunktur",
+  "settings.ai.localDiagnostic.models": "Líkön",
+  "settings.ai.localDiagnostic.tauriHttp": "Innbyggt Tauri HTTP",
+  "settings.ai.localDiagnostic.title": "Tengigreining",
+  "settings.ai.localDiagnostic.transport": "Flutningur",
   "settings.ai.ollamaBrowserNote": "Ollama er aðeins fáanlegt í skjáborðsforritinu. Innihaldsöryggisstefna vafrans lokar á beinar tengingar við localhost.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -1875,15 +1883,9 @@
   "settings.ai.providerStatusConnected": "Tengdur",
   "settings.ai.providerStatusDisconnected": "Ekki tiltækt",
   "settings.ai.providerStatusLabel": "Tengingarstaða",
-  "settings.ai.providerStatusReady": "Tilbúið",
   "settings.ai.providerStatusNotTested": "Ekki prófað",
+  "settings.ai.providerStatusReady": "Tilbúið",
   "settings.ai.providerStatusTesting": "Prófar…",
-  "settings.ai.localDiagnostic.title": "Tengigreining",
-  "settings.ai.localDiagnostic.endpoint": "Endapunktur",
-  "settings.ai.localDiagnostic.transport": "Flutningur",
-  "settings.ai.localDiagnostic.models": "Líkön",
-  "settings.ai.localDiagnostic.tauriHttp": "Innbyggt Tauri HTTP",
-  "settings.ai.localDiagnostic.browserFetch": "Vafrafyrirspurn",
   "settings.ai.providerStatusUnavailableBrowser": "Ekki í boði í vafra",
   "settings.ai.providerTitle": "AI veitandi",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -1906,8 +1908,8 @@
   "settings.ai.testError.desktopRequired": "Ollama og staðbundnir OpenAI-samhæfir netþjónar eru aðeins tiltækir í skjáborðsforritinu. Vafrar loka fyrir bein tengsl frá vefsíðum við localhost (CORS og Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "Enginn {{provider}} API lykill stilltur",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU er ekki tiltækt í þessum vafra — WebLLM þarfnast WebGPU (prófaðu Chrome/Edge eða virkjaðu flögg).",
   "settings.ai.testError.pluginUnavailable": "Staðbundið netþjónasamband ekki tiltækt: ekki tókst að hlaða HTTP-viðbót skjáborðs.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -1991,14 +1993,14 @@
   "settings.appearance.presetDefault": "Sjálfgefið",
   "settings.appearance.presetHint": "Forstillir lag á dökkt/ljóst val þitt. Writer's Sepia aðlagast að fullu að báðum stillingum.",
   "settings.appearance.presetSepia": "Sepia rithöfundar",
-  "settings.appearance.writingSurface": "Writing surface",
-  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
-  "settings.appearance.writingSurfaceTextured": "Textured",
-  "settings.appearance.writingSurfacePlain": "Plain",
   "settings.appearance.primaryColor": "Aðal litur",
   "settings.appearance.secondaryColor": "Secondary Litur",
   "settings.appearance.theme": "Umsóknarþema",
   "settings.appearance.title": "Útlit",
+  "settings.appearance.writingSurface": "Writing surface",
+  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
+  "settings.appearance.writingSurfacePlain": "Plain",
+  "settings.appearance.writingSurfaceTextured": "Textured",
   "settings.categories.about": "Um",
   "settings.categories.accessibility": "Aðgengi",
   "settings.categories.advancedAi": "Háþróuð gervigreind",
@@ -2335,13 +2337,13 @@
   "settings.privacy.encryptionSetAction": "Stilltu lykilorð",
   "settings.privacy.encryptionSetButton": "Virkja og dulkóða",
   "settings.privacy.encryptionSetup": "Stilltu lykilorð til að dulkóða öll handrit, skyndimyndir og stillingar sem vistaðar eru á þessu tæki.",
+  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.encryptionTooManyAttempts": "Of margar misheppnaðar tilraunir. Vinsamlegast bíddu í {{seconds}} sekúndur.",
   "settings.privacy.encryptionTooShort": "Aðgangsorð verður að vera að minnsta kosti 8 stafir",
   "settings.privacy.encryptionUnlockAction": "Opnaðu geymslu",
   "settings.privacy.encryptionUnlockButton": "Opnaðu",
   "settings.privacy.encryptionWarning": "Ef þú gleymir lykilorðinu þínu er ekki hægt að endurheimta gögnin þín.",
   "settings.privacy.encryptionWrongPassphrase": "Rangt lykilorð — afkóðun mistókst",
-  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.euDataResidency": "EU Gagnavistarleyfi",
   "settings.privacy.euDataResidencyHint": "Geymdu öll gögn innan netþjóna Evrópusambandsins þar sem við á.",
   "settings.privacy.localStorageOnly": "Aðeins staðbundin geymsla",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "Sækir {{model}} líkan ({{size}} MB). Þetta gerir staðbundna raddvinnslu kleift án skýjaleiðar.",
   "voice.modelDownload.error": "Niðurhal mistókst: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% lokið",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Reyndu aftur",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Raddlíkan niðurhal",
   "sidebar.characterGraph": "Persónurit",
   "sidebar.characterInterviews": "Persónuviðtöl",
diff --git a/public/locales/it/bundle.json b/public/locales/it/bundle.json
index b9a3f983..1f284731 100644
--- a/public/locales/it/bundle.json
+++ b/public/locales/it/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "Download annullato",
   "settings.ai.localAi.downloadEta": "circa {{seconds}} s rimanenti",
   "settings.ai.localAi.downloadFailed": "Download non riuscito. Riprova.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB di {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Download del modello IA in corso…",
   "settings.ai.localAi.downloadedLabel": "scaricato",
   "settings.ai.localAi.downloadingButton": "Download in corso…",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "Download del modello {{model}} ({{size}} MB). Questo abilita l'elaborazione vocale locale senza instradamento cloud.",
   "voice.modelDownload.error": "Download fallito: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% completato",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB di {{total}} MB",
   "voice.modelDownload.retry": "Riprova",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Download modello vocale",
   "sidebar.characterGraph": "Grafo dei personaggi",
   "sidebar.characterInterviews": "Interviste ai personaggi",
diff --git a/public/locales/ja/bundle.json b/public/locales/ja/bundle.json
index 9b8b0738..dff83140 100644
--- a/public/locales/ja/bundle.json
+++ b/public/locales/ja/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "ダウンロードがキャンセルされました",
   "settings.ai.localAi.downloadEta": "約。残り {{seconds}}秒",
   "settings.ai.localAi.downloadFailed": "ダウンロードに失敗しました。もう一度お試しください。",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "AI モデルをダウンロードしています…",
   "settings.ai.localAi.downloadedLabel": "ダウンロードされた",
   "settings.ai.localAi.downloadingButton": "ダウンロード中…",
@@ -1844,6 +1846,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "利用不可",
   "settings.ai.localBackendPreset": "ローカル バックエンド プリセット",
+  "settings.ai.localDiagnostic.browserFetch": "ブラウザー取得",
+  "settings.ai.localDiagnostic.endpoint": "エンドポイント",
+  "settings.ai.localDiagnostic.models": "モデル",
+  "settings.ai.localDiagnostic.tauriHttp": "Tauri ネイティブ HTTP",
+  "settings.ai.localDiagnostic.title": "接続診断",
+  "settings.ai.localDiagnostic.transport": "通信方式",
   "settings.ai.ollamaBrowserNote": "Ollama はデスクトップ アプリでのみ利用できます。ブラウザのコンテンツ セキュリティ ポリシーは、ローカルホストへの直接接続をブロックします。",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -1875,15 +1883,9 @@
   "settings.ai.providerStatusConnected": "接続済み",
   "settings.ai.providerStatusDisconnected": "利用不可",
   "settings.ai.providerStatusLabel": "接続ステータス",
-  "settings.ai.providerStatusReady": "準備ができて",
   "settings.ai.providerStatusNotTested": "未テスト",
+  "settings.ai.providerStatusReady": "準備ができて",
   "settings.ai.providerStatusTesting": "テスト中…",
-  "settings.ai.localDiagnostic.title": "接続診断",
-  "settings.ai.localDiagnostic.endpoint": "エンドポイント",
-  "settings.ai.localDiagnostic.transport": "通信方式",
-  "settings.ai.localDiagnostic.models": "モデル",
-  "settings.ai.localDiagnostic.tauriHttp": "Tauri ネイティブ HTTP",
-  "settings.ai.localDiagnostic.browserFetch": "ブラウザー取得",
   "settings.ai.providerStatusUnavailableBrowser": "ブラウザでは利用できません",
   "settings.ai.providerTitle": "AIプロバイダー",
   "settings.ai.providerTransformers": "トランスフォーマー.js",
@@ -1906,8 +1908,8 @@
   "settings.ai.testError.desktopRequired": "Ollama およびローカルの OpenAI 互換サーバーはデスクトップアプリでのみ利用できます。ブラウザは Web ページから localhost への直接接続をブロックします（CORS および Private Network Access）。",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "{{provider}} の API キーが設定されていません",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "このブラウザでは WebGPU が利用できません — WebLLM には WebGPU が必要です（Chrome/Edge を試すかフラグを有効にしてください）。",
   "settings.ai.testError.pluginUnavailable": "ローカルサーバーへの接続が利用できません: デスクトップ HTTP プラグインの読み込みに失敗しました。",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -1991,14 +1993,14 @@
   "settings.appearance.presetDefault": "デフォルト",
   "settings.appearance.presetHint": "プリセットはライト/ダークモードと組み合わさります。作家のセピアは両方のモードに完全対応しています。",
   "settings.appearance.presetSepia": "作家のセピア",
-  "settings.appearance.writingSurface": "執筆背景",
-  "settings.appearance.writingSurfaceHint": "長文執筆に最も心地よい背景スタイルを選んでください。",
-  "settings.appearance.writingSurfaceTextured": "テクスチャあり",
-  "settings.appearance.writingSurfacePlain": "無地",
   "settings.appearance.primaryColor": "原色",
   "settings.appearance.secondaryColor": "二次色",
   "settings.appearance.theme": "アプリケーションのテーマ",
   "settings.appearance.title": "外観",
+  "settings.appearance.writingSurface": "執筆背景",
+  "settings.appearance.writingSurfaceHint": "長文執筆に最も心地よい背景スタイルを選んでください。",
+  "settings.appearance.writingSurfacePlain": "無地",
+  "settings.appearance.writingSurfaceTextured": "テクスチャあり",
   "settings.categories.about": "について",
   "settings.categories.accessibility": "アクセシビリティ",
   "settings.categories.advancedAi": "高度なAI",
@@ -2335,13 +2337,13 @@
   "settings.privacy.encryptionSetAction": "パスフレーズを設定する",
   "settings.privacy.encryptionSetButton": "有効化と暗号化",
   "settings.privacy.encryptionSetup": "このデバイスに保存されているすべての原稿、スナップショット、設定を暗号化するためのパスフレーズを設定します。",
+  "settings.privacy.encryptionSetupFailed": "暗号化の設定を保存できませんでした — このブラウザでローカルストレージが許可されているか確認してから、もう一度お試しください",
   "settings.privacy.encryptionTooManyAttempts": "失敗した試行が多すぎます。 {{seconds}}秒待ってください。",
   "settings.privacy.encryptionTooShort": "パスフレーズは 8 文字以上である必要があります",
   "settings.privacy.encryptionUnlockAction": "ストレージのロックを解除する",
   "settings.privacy.encryptionUnlockButton": "ロックを解除する",
   "settings.privacy.encryptionWarning": "パスフレーズを忘れた場合、データは復元できません。",
   "settings.privacy.encryptionWrongPassphrase": "パスフレーズが間違っています - 復号化に失敗しました",
-  "settings.privacy.encryptionSetupFailed": "暗号化の設定を保存できませんでした — このブラウザでローカルストレージが許可されているか確認してから、もう一度お試しください",
   "settings.privacy.euDataResidency": "EUデータレジデンシー",
   "settings.privacy.euDataResidencyHint": "可能な限り、すべてのデータをEUサーバーに保持します。",
   "settings.privacy.localStorageOnly": "ローカルストレージのみ",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "{{model}} モデルをダウンロードしています ({{size}} MB)。これにより、クラウド ルーティングを使用せずにローカル音声処理が可能になります。",
   "voice.modelDownload.error": "ダウンロードに失敗しました: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% 完了",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "リトライ",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "音声モデルのダウンロード",
   "sidebar.characterGraph": "キャラクターグラフ",
   "sidebar.characterInterviews": "キャラクターインタビュー",
diff --git a/public/locales/ko/bundle.json b/public/locales/ko/bundle.json
index 4cae2b7b..aeec771d 100644
--- a/public/locales/ko/bundle.json
+++ b/public/locales/ko/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "다운로드가 취소되었습니다.",
   "settings.ai.localAi.downloadEta": "대략. {{seconds}}초 남음",
   "settings.ai.localAi.downloadFailed": "다운로드에 실패했습니다. 다시 시도해 주세요.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "AI 모델 다운로드 중…",
   "settings.ai.localAi.downloadedLabel": "다운로드됨",
   "settings.ai.localAi.downloadingButton": "다운로드 중…",
@@ -1844,6 +1846,12 @@
   "settings.ai.localAi.webgpuLabel": "웹GPU",
   "settings.ai.localAi.webgpuUnavailable": "사용할 수 없음",
   "settings.ai.localBackendPreset": "로컬 백엔드 사전 설정",
+  "settings.ai.localDiagnostic.browserFetch": "브라우저 가져오기",
+  "settings.ai.localDiagnostic.endpoint": "엔드포인트",
+  "settings.ai.localDiagnostic.models": "모델",
+  "settings.ai.localDiagnostic.tauriHttp": "Tauri 네이티브 HTTP",
+  "settings.ai.localDiagnostic.title": "연결 진단",
+  "settings.ai.localDiagnostic.transport": "전송 방식",
   "settings.ai.ollamaBrowserNote": "Ollama는 데스크톱 앱에서만 사용할 수 있습니다. 브라우저의 콘텐츠 보안 정책은 localhost에 대한 직접 연결을 차단합니다.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -1875,15 +1883,9 @@
   "settings.ai.providerStatusConnected": "연결됨",
   "settings.ai.providerStatusDisconnected": "없는",
   "settings.ai.providerStatusLabel": "연결 상태",
-  "settings.ai.providerStatusReady": "준비가 된",
   "settings.ai.providerStatusNotTested": "테스트되지 않음",
+  "settings.ai.providerStatusReady": "준비가 된",
   "settings.ai.providerStatusTesting": "테스트 중…",
-  "settings.ai.localDiagnostic.title": "연결 진단",
-  "settings.ai.localDiagnostic.endpoint": "엔드포인트",
-  "settings.ai.localDiagnostic.transport": "전송 방식",
-  "settings.ai.localDiagnostic.models": "모델",
-  "settings.ai.localDiagnostic.tauriHttp": "Tauri 네이티브 HTTP",
-  "settings.ai.localDiagnostic.browserFetch": "브라우저 가져오기",
   "settings.ai.providerStatusUnavailableBrowser": "브라우저에서 사용할 수 없음",
   "settings.ai.providerTitle": "AI 제공업체",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -1906,8 +1908,8 @@
   "settings.ai.testError.desktopRequired": "Ollama 및 로컬 OpenAI 호환 서버는 데스크톱 앱에서만 사용할 수 있습니다. 브라우저는 웹 페이지에서 localhost로의 직접 연결을 차단합니다(CORS 및 Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "{{provider}} API 키가 설정되지 않았습니다",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "이 브라우저에서는 WebGPU를 사용할 수 없습니다 — WebLLM에는 WebGPU가 필요합니다(Chrome/Edge를 사용하거나 플래그를 활성화하세요).",
   "settings.ai.testError.pluginUnavailable": "로컬 서버 네트워킹을 사용할 수 없습니다: 데스크톱 HTTP 플러그인을 불러오지 못했습니다.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -1991,14 +1993,14 @@
   "settings.appearance.presetDefault": "기본",
   "settings.appearance.presetHint": "어둡거나 밝은 선택으로 레이어를 사전 설정합니다. 작가의 세피아는 두 모드 모두에 완전히 적응합니다.",
   "settings.appearance.presetSepia": "작가의 세피아",
-  "settings.appearance.writingSurface": "글쓰기 배경",
-  "settings.appearance.writingSurfaceHint": "장문 집필에 가장 편안한 배경 스타일을 선택하세요.",
-  "settings.appearance.writingSurfaceTextured": "질감 있음",
-  "settings.appearance.writingSurfacePlain": "단색",
   "settings.appearance.primaryColor": "원색",
   "settings.appearance.secondaryColor": "보조 색상",
   "settings.appearance.theme": "응용 테마",
   "settings.appearance.title": "모습",
+  "settings.appearance.writingSurface": "글쓰기 배경",
+  "settings.appearance.writingSurfaceHint": "장문 집필에 가장 편안한 배경 스타일을 선택하세요.",
+  "settings.appearance.writingSurfacePlain": "단색",
+  "settings.appearance.writingSurfaceTextured": "질감 있음",
   "settings.categories.about": "에 대한",
   "settings.categories.accessibility": "접근성",
   "settings.categories.advancedAi": "고급 AI",
@@ -2335,13 +2337,13 @@
   "settings.privacy.encryptionSetAction": "암호 설정",
   "settings.privacy.encryptionSetButton": "활성화 및 암호화",
   "settings.privacy.encryptionSetup": "이 장치에 저장된 모든 원고, 스냅샷 및 설정을 암호화하려면 암호를 설정하세요.",
+  "settings.privacy.encryptionSetupFailed": "암호화 설정을 저장할 수 없습니다 — 이 브라우저가 로컬 저장소를 허용하는지 확인한 후 다시 시도하세요",
   "settings.privacy.encryptionTooManyAttempts": "실패한 시도가 너무 많습니다. {{seconds}}초만 기다려주세요.",
   "settings.privacy.encryptionTooShort": "암호는 8자 이상이어야 합니다.",
   "settings.privacy.encryptionUnlockAction": "스토리지 잠금 해제",
   "settings.privacy.encryptionUnlockButton": "터놓다",
   "settings.privacy.encryptionWarning": "암호를 잊어버리면 데이터를 복구할 수 없습니다.",
   "settings.privacy.encryptionWrongPassphrase": "잘못된 암호 - 암호 해독 실패",
-  "settings.privacy.encryptionSetupFailed": "암호화 설정을 저장할 수 없습니다 — 이 브라우저가 로컬 저장소를 허용하는지 확인한 후 다시 시도하세요",
   "settings.privacy.euDataResidency": "EU 데이터 상주",
   "settings.privacy.euDataResidencyHint": "해당하는 경우 모든 데이터를 유럽 연합 서버 내에 보관하십시오.",
   "settings.privacy.localStorageOnly": "로컬 저장소만",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "{{model}} 모델 다운로드 중({{size}}MB) 이를 통해 클라우드 라우팅 없이 로컬 음성 처리가 가능해집니다.",
   "voice.modelDownload.error": "다운로드 실패: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% 완료",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "다시 해 보다",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "음성 모델 다운로드",
   "sidebar.characterGraph": "캐릭터 그래프",
   "sidebar.characterInterviews": "캐릭터 인터뷰",
diff --git a/public/locales/pt/bundle.json b/public/locales/pt/bundle.json
index df71dc8d..428b7816 100644
--- a/public/locales/pt/bundle.json
+++ b/public/locales/pt/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "Download cancelado",
   "settings.ai.localAi.downloadEta": "aprox. {{seconds}} restantes",
   "settings.ai.localAi.downloadFailed": "A transferência falhou. Tente novamente.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Downloading IA model…",
   "settings.ai.localAi.downloadedLabel": "baixado",
   "settings.ai.localAi.downloadingButton": "A transferir…",
@@ -1844,6 +1846,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Indisponível",
   "settings.ai.localBackendPreset": "Predefinição de back-end local",
+  "settings.ai.localDiagnostic.browserFetch": "Solicitação do navegador",
+  "settings.ai.localDiagnostic.endpoint": "Ponto de extremidade",
+  "settings.ai.localDiagnostic.models": "Modelos",
+  "settings.ai.localDiagnostic.tauriHttp": "HTTP nativo do Tauri",
+  "settings.ai.localDiagnostic.title": "Diagnóstico de conexão",
+  "settings.ai.localDiagnostic.transport": "Transporte",
   "settings.ai.ollamaBrowserNote": "Ollama está disponível apenas no aplicativo de desktop. A Política de Segurança de Conteúdo do navegador bloqueia conexões diretas com o host local.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -1875,15 +1883,9 @@
   "settings.ai.providerStatusConnected": "Conectado",
   "settings.ai.providerStatusDisconnected": "Indisponível",
   "settings.ai.providerStatusLabel": "Status da conexão",
-  "settings.ai.providerStatusReady": "Preparar",
   "settings.ai.providerStatusNotTested": "Não testado",
+  "settings.ai.providerStatusReady": "Preparar",
   "settings.ai.providerStatusTesting": "Testando…",
-  "settings.ai.localDiagnostic.title": "Diagnóstico de conexão",
-  "settings.ai.localDiagnostic.endpoint": "Ponto de extremidade",
-  "settings.ai.localDiagnostic.transport": "Transporte",
-  "settings.ai.localDiagnostic.models": "Modelos",
-  "settings.ai.localDiagnostic.tauriHttp": "HTTP nativo do Tauri",
-  "settings.ai.localDiagnostic.browserFetch": "Solicitação do navegador",
   "settings.ai.providerStatusUnavailableBrowser": "Não disponível no navegador",
   "settings.ai.providerTitle": "IA Provider",
   "settings.ai.providerTransformers": "Transformadores.js",
@@ -1906,8 +1908,8 @@
   "settings.ai.testError.desktopRequired": "O Ollama e os servidores locais compatíveis com OpenAI só estão disponíveis na aplicação de ambiente de trabalho. Os navegadores bloqueiam ligações diretas de páginas web ao localhost (CORS e Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "O servidor local devolveu uma resposta de modelos inválida.",
-  "settings.ai.testError.noModels": "O servidor local está acessível, mas não expõe modelos.",
   "settings.ai.testError.noApiKey": "Nenhuma chave de API {{provider}} definida",
+  "settings.ai.testError.noModels": "O servidor local está acessível, mas não expõe modelos.",
   "settings.ai.testError.noWebgpu": "WebGPU indisponível neste navegador — o WebLLM necessita de WebGPU (experimente o Chrome/Edge ou ative as flags).",
   "settings.ai.testError.pluginUnavailable": "A rede do servidor local não está disponível: o plugin HTTP de ambiente de trabalho falhou ao carregar.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -1991,14 +1993,14 @@
   "settings.appearance.presetDefault": "Padrão",
   "settings.appearance.presetHint": "Os presets se combinam com seu modo claro/escuro. O Sépia do escritor adapta-se totalmente a ambos os modos.",
   "settings.appearance.presetSepia": "Sépia do escritor",
-  "settings.appearance.writingSurface": "Superfície de escrita",
-  "settings.appearance.writingSurfaceHint": "Escolha o estilo de fundo mais confortável para escrita longa.",
-  "settings.appearance.writingSurfaceTextured": "Texturizada",
-  "settings.appearance.writingSurfacePlain": "Lisa",
   "settings.appearance.primaryColor": "Cor Primária",
   "settings.appearance.secondaryColor": "Cor Secundária",
   "settings.appearance.theme": "Tema do aplicativo",
   "settings.appearance.title": "Aparência",
+  "settings.appearance.writingSurface": "Superfície de escrita",
+  "settings.appearance.writingSurfaceHint": "Escolha o estilo de fundo mais confortável para escrita longa.",
+  "settings.appearance.writingSurfacePlain": "Lisa",
+  "settings.appearance.writingSurfaceTextured": "Texturizada",
   "settings.categories.about": "Sobre",
   "settings.categories.accessibility": "Acessibilidade",
   "settings.categories.advancedAi": "Advanced IA",
@@ -2335,13 +2337,13 @@
   "settings.privacy.encryptionSetAction": "Definir senha",
   "settings.privacy.encryptionSetButton": "Habilitar e criptografar",
   "settings.privacy.encryptionSetup": "Defina uma senha para criptografar todos os manuscritos, instantâneos e configurações armazenados neste dispositivo.",
+  "settings.privacy.encryptionSetupFailed": "Não foi possível salvar a configuração de criptografia — verifique se este navegador permite armazenamento local e tente novamente",
   "settings.privacy.encryptionTooManyAttempts": "Muitas tentativas fracassadas. Aguarde {{seconds}} segundos.",
   "settings.privacy.encryptionTooShort": "A senha deve ter pelo menos 8 caracteres",
   "settings.privacy.encryptionUnlockAction": "Desbloquear armazenamento",
   "settings.privacy.encryptionUnlockButton": "Desbloquear",
   "settings.privacy.encryptionWarning": "Se você esquecer sua senha, seus dados não poderão ser recuperados.",
   "settings.privacy.encryptionWrongPassphrase": "Senha errada – falha na descriptografia",
-  "settings.privacy.encryptionSetupFailed": "Não foi possível salvar a configuração de criptografia — verifique se este navegador permite armazenamento local e tente novamente",
   "settings.privacy.euDataResidency": "Residência de Dados na UE",
   "settings.privacy.euDataResidencyHint": "Manter todos os dados em servidores da União Europeia quando possível.",
   "settings.privacy.localStorageOnly": "Somente armazenamento local",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "Fazendo download do modelo {{model}} ({{size}} MB). Isso permite o processamento de voz local sem roteamento na nuvem.",
   "voice.modelDownload.error": "Falha no download: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% concluído",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Tentar novamente",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Download do modelo de voz",
   "sidebar.characterGraph": "Gráfico de caracteres",
   "sidebar.characterInterviews": "Entrevistas com personagens",
diff --git a/public/locales/ru/bundle.json b/public/locales/ru/bundle.json
index 950646aa..feacb2da 100644
--- a/public/locales/ru/bundle.json
+++ b/public/locales/ru/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "Загрузка отменена",
   "settings.ai.localAi.downloadEta": "ок. осталось {{seconds}}с",
   "settings.ai.localAi.downloadFailed": "Загрузка не удалась. Пожалуйста, попробуйте еще раз.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Загрузка модели ИИ…",
   "settings.ai.localAi.downloadedLabel": "скачал",
   "settings.ai.localAi.downloadingButton": "Загрузка…",
@@ -1844,6 +1846,12 @@
   "settings.ai.localAi.webgpuLabel": "Веб-графический процессор",
   "settings.ai.localAi.webgpuUnavailable": "Нет в наличии",
   "settings.ai.localBackendPreset": "Локальная настройка серверной части",
+  "settings.ai.localDiagnostic.browserFetch": "Запрос браузера",
+  "settings.ai.localDiagnostic.endpoint": "Конечная точка",
+  "settings.ai.localDiagnostic.models": "Модели",
+  "settings.ai.localDiagnostic.tauriHttp": "Нативный HTTP Tauri",
+  "settings.ai.localDiagnostic.title": "Диагностика подключения",
+  "settings.ai.localDiagnostic.transport": "Транспорт",
   "settings.ai.ollamaBrowserNote": "Оллама доступен только в настольном приложении. Политика безопасности контента браузера блокирует прямые подключения к локальному хосту.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -1875,15 +1883,9 @@
   "settings.ai.providerStatusConnected": "Подключено",
   "settings.ai.providerStatusDisconnected": "Недоступно",
   "settings.ai.providerStatusLabel": "Статус соединения",
-  "settings.ai.providerStatusReady": "Готовый",
   "settings.ai.providerStatusNotTested": "Не проверено",
+  "settings.ai.providerStatusReady": "Готовый",
   "settings.ai.providerStatusTesting": "Проверка…",
-  "settings.ai.localDiagnostic.title": "Диагностика подключения",
-  "settings.ai.localDiagnostic.endpoint": "Конечная точка",
-  "settings.ai.localDiagnostic.transport": "Транспорт",
-  "settings.ai.localDiagnostic.models": "Модели",
-  "settings.ai.localDiagnostic.tauriHttp": "Нативный HTTP Tauri",
-  "settings.ai.localDiagnostic.browserFetch": "Запрос браузера",
   "settings.ai.providerStatusUnavailableBrowser": "Недоступно в браузере",
   "settings.ai.providerTitle": "Поставщик ИИ",
   "settings.ai.providerTransformers": "Трансформеры.js",
@@ -1906,8 +1908,8 @@
   "settings.ai.testError.desktopRequired": "Ollama и локальные серверы, совместимые с OpenAI, доступны только в настольном приложении. Браузеры блокируют прямые подключения с веб-страниц к localhost (CORS и Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "Не задан ключ API {{provider}}",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU недоступен в этом браузере — WebLLM требует WebGPU (попробуйте Chrome/Edge или включите флаги).",
   "settings.ai.testError.pluginUnavailable": "Сеть локального сервера недоступна: не удалось загрузить настольный HTTP-плагин.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -1991,14 +1993,14 @@
   "settings.appearance.presetDefault": "По умолчанию",
   "settings.appearance.presetHint": "Слой предустановок по вашему выбору темный/светлый. Writer's Sepia полностью адаптируется к обоим режимам.",
   "settings.appearance.presetSepia": "Писательская сепия",
-  "settings.appearance.writingSurface": "Фон для письма",
-  "settings.appearance.writingSurfaceHint": "Выберите стиль фона, который наиболее комфортен для долгой работы с текстом.",
-  "settings.appearance.writingSurfaceTextured": "Текстурный",
-  "settings.appearance.writingSurfacePlain": "Однотонный",
   "settings.appearance.primaryColor": "Основной цвет",
   "settings.appearance.secondaryColor": "Вторичный цвет",
   "settings.appearance.theme": "Тема приложения",
   "settings.appearance.title": "Появление",
+  "settings.appearance.writingSurface": "Фон для письма",
+  "settings.appearance.writingSurfaceHint": "Выберите стиль фона, который наиболее комфортен для долгой работы с текстом.",
+  "settings.appearance.writingSurfacePlain": "Однотонный",
+  "settings.appearance.writingSurfaceTextured": "Текстурный",
   "settings.categories.about": "О",
   "settings.categories.accessibility": "Доступность",
   "settings.categories.advancedAi": "Продвинутый ИИ",
@@ -2335,13 +2337,13 @@
   "settings.privacy.encryptionSetAction": "Установить парольную фразу",
   "settings.privacy.encryptionSetButton": "Включить и зашифровать",
   "settings.privacy.encryptionSetup": "Установите парольную фразу для шифрования всех рукописей, снимков и настроек, хранящихся на этом устройстве.",
+  "settings.privacy.encryptionSetupFailed": "Не удалось сохранить настройку шифрования — убедитесь, что этот браузер разрешает локальное хранилище, и повторите попытку",
   "settings.privacy.encryptionTooManyAttempts": "Слишком много неудачных попыток. Пожалуйста, подождите {{seconds}} секунд.",
   "settings.privacy.encryptionTooShort": "Парольная фраза должна содержать не менее 8 символов.",
   "settings.privacy.encryptionUnlockAction": "Разблокировать хранилище",
   "settings.privacy.encryptionUnlockButton": "Разблокировать",
   "settings.privacy.encryptionWarning": "Если вы забудете парольную фразу, ваши данные не смогут быть восстановлены.",
   "settings.privacy.encryptionWrongPassphrase": "Неверная парольная фраза — расшифровка не удалась.",
-  "settings.privacy.encryptionSetupFailed": "Не удалось сохранить настройку шифрования — убедитесь, что этот браузер разрешает локальное хранилище, и повторите попытку",
   "settings.privacy.euDataResidency": "Резиденция данных ЕС",
   "settings.privacy.euDataResidencyHint": "Храните все данные на серверах Европейского Союза, где это применимо.",
   "settings.privacy.localStorageOnly": "Только локальное хранилище",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "Скачиваем модель {{model}} ({{size}} МБ). Это обеспечивает локальную обработку голоса без облачной маршрутизации.",
   "voice.modelDownload.error": "Загрузка не удалась: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% выполнено",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Повторить попытку",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Скачать модель голоса",
   "sidebar.characterGraph": "График персонажей",
   "sidebar.characterInterviews": "Интервью с персонажами",
diff --git a/public/locales/sv/bundle.json b/public/locales/sv/bundle.json
index 33ec166b..c50e0409 100644
--- a/public/locales/sv/bundle.json
+++ b/public/locales/sv/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "Nedladdningen avbröts",
   "settings.ai.localAi.downloadEta": "ca. {{seconds}}s kvar",
   "settings.ai.localAi.downloadFailed": "Nedladdningen misslyckades. Försök igen.",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Laddar ner AI-modell...",
   "settings.ai.localAi.downloadedLabel": "nedladdade",
   "settings.ai.localAi.downloadingButton": "Laddar ned...",
@@ -1844,6 +1846,12 @@
   "settings.ai.localAi.webgpuLabel": "WebGPU",
   "settings.ai.localAi.webgpuUnavailable": "Ej tillgängligt",
   "settings.ai.localBackendPreset": "Lokal förinställning av backend",
+  "settings.ai.localDiagnostic.browserFetch": "Webbläsarhämtning",
+  "settings.ai.localDiagnostic.endpoint": "Slutpunkt",
+  "settings.ai.localDiagnostic.models": "Modeller",
+  "settings.ai.localDiagnostic.tauriHttp": "Tauris inbyggda HTTP",
+  "settings.ai.localDiagnostic.title": "Anslutningsdiagnostik",
+  "settings.ai.localDiagnostic.transport": "Transport",
   "settings.ai.ollamaBrowserNote": "Ollama är endast tillgänglig i skrivbordsappen. Webbläsarens innehållssäkerhetspolicy blockerar direktanslutningar till localhost.",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -1875,15 +1883,9 @@
   "settings.ai.providerStatusConnected": "Ansluten",
   "settings.ai.providerStatusDisconnected": "Inte tillgänglig",
   "settings.ai.providerStatusLabel": "Anslutningsstatus",
-  "settings.ai.providerStatusReady": "Redo",
   "settings.ai.providerStatusNotTested": "Inte testad",
+  "settings.ai.providerStatusReady": "Redo",
   "settings.ai.providerStatusTesting": "Testar…",
-  "settings.ai.localDiagnostic.title": "Anslutningsdiagnostik",
-  "settings.ai.localDiagnostic.endpoint": "Slutpunkt",
-  "settings.ai.localDiagnostic.transport": "Transport",
-  "settings.ai.localDiagnostic.models": "Modeller",
-  "settings.ai.localDiagnostic.tauriHttp": "Tauris inbyggda HTTP",
-  "settings.ai.localDiagnostic.browserFetch": "Webbläsarhämtning",
   "settings.ai.providerStatusUnavailableBrowser": "Inte tillgängligt i webbläsaren",
   "settings.ai.providerTitle": "AI-leverantör",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -1906,8 +1908,8 @@
   "settings.ai.testError.desktopRequired": "Ollama och lokala OpenAI-kompatibla servrar är endast tillgängliga i skrivbordsappen. Webbläsare blockerar direkta anslutningar från webbsidor till localhost (CORS och Private Network Access).",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "Ingen {{provider}}-API-nyckel angiven",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "WebGPU är inte tillgängligt i den här webbläsaren — WebLLM behöver WebGPU (prova Chrome/Edge eller aktivera flaggor).",
   "settings.ai.testError.pluginUnavailable": "Lokal servernätverksanslutning ej tillgänglig: skrivbordets HTTP-plugin misslyckades med att läsas in.",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -1991,14 +1993,14 @@
   "settings.appearance.presetDefault": "Standard",
   "settings.appearance.presetHint": "Förinställer lager på ditt mörka/ljusa val. Writer's Sepia anpassar sig helt till båda lägena.",
   "settings.appearance.presetSepia": "Författarens Sepia",
-  "settings.appearance.writingSurface": "Writing surface",
-  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
-  "settings.appearance.writingSurfaceTextured": "Textured",
-  "settings.appearance.writingSurfacePlain": "Plain",
   "settings.appearance.primaryColor": "Grundfärg",
   "settings.appearance.secondaryColor": "Sekundär färg",
   "settings.appearance.theme": "Programtema",
   "settings.appearance.title": "Utseende",
+  "settings.appearance.writingSurface": "Writing surface",
+  "settings.appearance.writingSurfaceHint": "Pick the backdrop style that feels most comfortable for long-form writing.",
+  "settings.appearance.writingSurfacePlain": "Plain",
+  "settings.appearance.writingSurfaceTextured": "Textured",
   "settings.categories.about": "Om",
   "settings.categories.accessibility": "Tillgänglighet",
   "settings.categories.advancedAi": "Avancerad AI",
@@ -2335,13 +2337,13 @@
   "settings.privacy.encryptionSetAction": "Ställ in lösenordsfras",
   "settings.privacy.encryptionSetButton": "Aktivera och kryptera",
   "settings.privacy.encryptionSetup": "Ställ in en lösenordsfras för att kryptera alla manuskript, ögonblicksbilder och inställningar som lagras på den här enheten.",
+  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.encryptionTooManyAttempts": "För många misslyckade försök. Vänta {{seconds}} sekunder.",
   "settings.privacy.encryptionTooShort": "Lösenfrasen måste bestå av minst 8 tecken",
   "settings.privacy.encryptionUnlockAction": "Lås upp lagring",
   "settings.privacy.encryptionUnlockButton": "Låsa upp",
   "settings.privacy.encryptionWarning": "Om du glömmer din lösenfras kan din data inte återställas.",
   "settings.privacy.encryptionWrongPassphrase": "Fel lösenfras — dekrypteringen misslyckades",
-  "settings.privacy.encryptionSetupFailed": "Could not save the encryption setup — check that this browser allows local storage, then try again",
   "settings.privacy.euDataResidency": "EU Data Residency",
   "settings.privacy.euDataResidencyHint": "Behåll all data på Europeiska unionens servrar där det är tillämpligt.",
   "settings.privacy.localStorageOnly": "Endast lokal lagring",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "Laddar ner {{model}}-modellen ({{size}} MB). Detta möjliggör lokal röstbehandling utan molnrouting.",
   "voice.modelDownload.error": "Nedladdning misslyckades: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% klar",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "Försöka igen",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Ladda ner röstmodell",
   "sidebar.characterGraph": "Karaktärsgraf",
   "sidebar.characterInterviews": "Karaktärsintervjuer",
diff --git a/public/locales/zh/bundle.json b/public/locales/zh/bundle.json
index 226d5ccd..976f25ba 100644
--- a/public/locales/zh/bundle.json
+++ b/public/locales/zh/bundle.json
@@ -1805,6 +1805,8 @@
   "settings.ai.localAi.downloadCancelled": "下载已取消",
   "settings.ai.localAi.downloadEta": "约。还剩 {{seconds}} 秒",
   "settings.ai.localAi.downloadFailed": "下载失败。请重试。",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "正在下载 AI 模型...",
   "settings.ai.localAi.downloadedLabel": "下载的",
   "settings.ai.localAi.downloadingButton": "下载中…",
@@ -1844,6 +1846,12 @@
   "settings.ai.localAi.webgpuLabel": "网络GPU",
   "settings.ai.localAi.webgpuUnavailable": "不可用",
   "settings.ai.localBackendPreset": "本地后端预设",
+  "settings.ai.localDiagnostic.browserFetch": "浏览器请求",
+  "settings.ai.localDiagnostic.endpoint": "端点",
+  "settings.ai.localDiagnostic.models": "模型",
+  "settings.ai.localDiagnostic.tauriHttp": "Tauri 原生 HTTP",
+  "settings.ai.localDiagnostic.title": "连接诊断",
+  "settings.ai.localDiagnostic.transport": "传输方式",
   "settings.ai.ollamaBrowserNote": "Ollama 仅在桌面应用程序中可用。浏览器的内容安全策略阻止与本地主机的直接连接。",
   "settings.ai.ollamaBrowserOptInBody": "Experimental — start your own Ollama server with the command below so it accepts requests from this exact page, then reload. This isn't a proxy or a bypass: it only works when you've configured it yourself.",
   "settings.ai.ollamaBrowserOptInTitle": "Browser connection enabled (advanced, unsupported)",
@@ -1875,15 +1883,9 @@
   "settings.ai.providerStatusConnected": "已连接",
   "settings.ai.providerStatusDisconnected": "不可用",
   "settings.ai.providerStatusLabel": "连接状态",
-  "settings.ai.providerStatusReady": "准备好",
   "settings.ai.providerStatusNotTested": "未测试",
+  "settings.ai.providerStatusReady": "准备好",
   "settings.ai.providerStatusTesting": "测试中…",
-  "settings.ai.localDiagnostic.title": "连接诊断",
-  "settings.ai.localDiagnostic.endpoint": "端点",
-  "settings.ai.localDiagnostic.transport": "传输方式",
-  "settings.ai.localDiagnostic.models": "模型",
-  "settings.ai.localDiagnostic.tauriHttp": "Tauri 原生 HTTP",
-  "settings.ai.localDiagnostic.browserFetch": "浏览器请求",
   "settings.ai.providerStatusUnavailableBrowser": "浏览器中不可用",
   "settings.ai.providerTitle": "人工智能提供商",
   "settings.ai.providerTransformers": "Transformers.js",
@@ -1906,8 +1908,8 @@
   "settings.ai.testError.desktopRequired": "Ollama 和本地 OpenAI 兼容服务器仅在桌面应用中可用。浏览器会阻止网页直接连接到 localhost（CORS 和 Private Network Access）。",
   "settings.ai.testError.httpError": "HTTP {{status}}",
   "settings.ai.testError.invalidResponse": "The local server returned an invalid models response.",
-  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noApiKey": "未设置 {{provider}} API 密钥",
+  "settings.ai.testError.noModels": "The local server is reachable but exposes no models.",
   "settings.ai.testError.noWebgpu": "此浏览器不支持 WebGPU — WebLLM 需要 WebGPU（请尝试 Chrome/Edge 或启用相关标志）。",
   "settings.ai.testError.pluginUnavailable": "本地服务器网络不可用：桌面 HTTP 插件加载失败。",
   "settings.ai.testError.proxyUnavailableStaticHost": "Claude is not available on this deployment: it has no serverless proxy (GitHub Pages is static-only). Use the desktop app, or a Vercel/Cloudflare Pages deployment instead.",
@@ -1991,14 +1993,14 @@
   "settings.appearance.presetDefault": "默认",
   "settings.appearance.presetHint": "预设与深色/浅色模式叠加。作家的棕褐色可完全适应两种模式。",
   "settings.appearance.presetSepia": "作家的棕褐色",
-  "settings.appearance.writingSurface": "写作背景",
-  "settings.appearance.writingSurfaceHint": "选择最适合长篇写作的背景样式。",
-  "settings.appearance.writingSurfaceTextured": "有纹理",
-  "settings.appearance.writingSurfacePlain": "纯色",
   "settings.appearance.primaryColor": "原色",
   "settings.appearance.secondaryColor": "次要颜色",
   "settings.appearance.theme": "应用主题",
   "settings.appearance.title": "外貌",
+  "settings.appearance.writingSurface": "写作背景",
+  "settings.appearance.writingSurfaceHint": "选择最适合长篇写作的背景样式。",
+  "settings.appearance.writingSurfacePlain": "纯色",
+  "settings.appearance.writingSurfaceTextured": "有纹理",
   "settings.categories.about": "关于",
   "settings.categories.accessibility": "无障碍",
   "settings.categories.advancedAi": "先进的人工智能",
@@ -2335,13 +2337,13 @@
   "settings.privacy.encryptionSetAction": "设置密码",
   "settings.privacy.encryptionSetButton": "启用和加密",
   "settings.privacy.encryptionSetup": "设置密码以加密存储在该设备上的所有手稿、快照和设置。",
+  "settings.privacy.encryptionSetupFailed": "无法保存加密设置 — 请检查此浏览器是否允许本地存储，然后重试",
   "settings.privacy.encryptionTooManyAttempts": "失败的尝试太多了。请等待{{seconds}}秒。",
   "settings.privacy.encryptionTooShort": "密码必须至少为 8 个字符",
   "settings.privacy.encryptionUnlockAction": "解锁存储",
   "settings.privacy.encryptionUnlockButton": "开锁",
   "settings.privacy.encryptionWarning": "如果您忘记密码，您的数据将无法恢复。",
   "settings.privacy.encryptionWrongPassphrase": "密码错误 — 解密失败",
-  "settings.privacy.encryptionSetupFailed": "无法保存加密设置 — 请检查此浏览器是否允许本地存储，然后重试",
   "settings.privacy.euDataResidency": "欧盟数据驻留",
   "settings.privacy.euDataResidencyHint": "在可能的情况下，将所有数据保存在欧盟服务器上。",
   "settings.privacy.localStorageOnly": "仅本地存储",
@@ -2465,7 +2467,9 @@
   "voice.modelDownload.description": "正在下载 {{model}} 模型 ({{size}} MB)。这使得本地语音处理无需云路由。",
   "voice.modelDownload.error": "下载失败：{{error}}",
   "voice.modelDownload.progress": "{{percent}}% 完成",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
   "voice.modelDownload.retry": "重试",
+  "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "语音模型下载",
   "sidebar.characterGraph": "人物图",
   "sidebar.characterInterviews": "人物访谈",
diff --git a/services/ai/inferenceProgressEmitter.ts b/services/ai/inferenceProgressEmitter.ts
index 1eea7b58..6aa21595 100644
--- a/services/ai/inferenceProgressEmitter.ts
+++ b/services/ai/inferenceProgressEmitter.ts
@@ -2,6 +2,8 @@
 //          the UI can subscribe without dispatching on every 1% progress tick.
 //          Adapted from CannaGuide-2025 progressEmitter.ts for WorldScript context.
 
+import { WEBLLM_MODEL_APPROX_MB, type WebLlmModelId } from '@domain/ai-core';
+
 export type WebLlmLoadingState = 'idle' | 'loading' | 'ready' | 'error';
 
 export interface WebLlmLoadProgress {
@@ -9,17 +11,27 @@ export interface WebLlmLoadProgress {
   progress: number; // 0–1
   text: string;
   estimatedSecondsRemaining: number | null;
+  // QNBS-v3 (#333 item 1): DERIVED, not measured — WebLLM's own progress callback exposes only a
+  // 0-1 fraction, no structured byte counts (verified against @mlc-ai/web-llm's InitProgressReport
+  // type). loadedBytes/totalBytes are progress × a hand-authored known-total-size table
+  // (WEBLLM_MODEL_APPROX_MB); null whenever the model id isn't in that table.
+  loadedBytes: number | null;
+  totalBytes: number | null;
+  bytesPerSecond: number | null;
 }
 
-type ProgressListener = (snapshot: WebLlmLoadProgress) => void;
-
 const INITIAL_SNAPSHOT: WebLlmLoadProgress = {
   state: 'idle',
   progress: 0,
   text: '',
   estimatedSecondsRemaining: null,
+  loadedBytes: null,
+  totalBytes: null,
+  bytesPerSecond: null,
 };
 
+type ProgressListener = (snapshot: WebLlmLoadProgress) => void;
+
 class InferenceProgressEmitter {
   private snapshot: WebLlmLoadProgress = { ...INITIAL_SNAPSHOT };
   private listeners: Set<ProgressListener> = new Set();
@@ -41,7 +53,9 @@ class InferenceProgressEmitter {
     return { ...this.snapshot };
   }
 
-  reportWebLlmProgress(progress: number, text: string): void {
+  // QNBS-v3 (#333 item 1): modelId is optional (existing callers with no known model id keep
+  // working; loadedBytes/totalBytes/bytesPerSecond simply stay null when it's absent or unknown).
+  reportWebLlmProgress(progress: number, text: string, modelId?: string): void {
     if (this.snapshot.state !== 'loading') {
       this.loadStartMs = Date.now();
     }
@@ -50,13 +64,35 @@ class InferenceProgressEmitter {
     const estimatedSecondsRemaining =
       progress > 0.01 && progress < 1 ? Math.round((elapsed / progress) * (1 - progress)) : null;
 
-    this.snapshot = { state: 'loading', progress, text, estimatedSecondsRemaining };
+    const approxMb = modelId ? WEBLLM_MODEL_APPROX_MB[modelId as WebLlmModelId] : undefined;
+    const totalBytes = approxMb != null ? approxMb * 1024 * 1024 : null;
+    const loadedBytes = totalBytes != null ? Math.round(totalBytes * progress) : null;
+    const bytesPerSecond =
+      loadedBytes != null && elapsed > 0.5 ? Math.round(loadedBytes / elapsed) : null;
+
+    this.snapshot = {
+      state: 'loading',
+      progress,
+      text,
+      estimatedSecondsRemaining,
+      loadedBytes,
+      totalBytes,
+      bytesPerSecond,
+    };
     this.emit();
   }
 
   reportWebLlmReady(): void {
     this.loadStartMs = null;
-    this.snapshot = { state: 'ready', progress: 1, text: '', estimatedSecondsRemaining: null };
+    this.snapshot = {
+      state: 'ready',
+      progress: 1,
+      text: '',
+      estimatedSecondsRemaining: null,
+      loadedBytes: null,
+      totalBytes: null,
+      bytesPerSecond: null,
+    };
     this.emit();
   }
 
@@ -67,6 +103,9 @@ class InferenceProgressEmitter {
       progress: this.snapshot.progress,
       text: message,
       estimatedSecondsRemaining: null,
+      loadedBytes: null,
+      totalBytes: null,
+      bytesPerSecond: null,
     };
     this.emit();
   }
diff --git a/services/downloadProgressFormat.ts b/services/downloadProgressFormat.ts
new file mode 100644
index 00000000..d45bab88
--- /dev/null
+++ b/services/downloadProgressFormat.ts
@@ -0,0 +1,13 @@
+// QNBS-v3 (#333 item 1): shared formatting for the two model-download progress UIs
+// (LocalAiDownloadProgress, VoiceModelDownloadModal) — kept as pure functions so both can render
+// identical "X MB of Y MB" / "Z MB/s" text without duplicating the rounding/formatting logic.
+
+/** Formats a byte count as whole megabytes, e.g. `formatMegabytes(43_000_000)` → `"41"`. */
+export function formatMegabytes(bytes: number): string {
+  return (bytes / (1024 * 1024)).toFixed(0);
+}
+
+/** Formats a bytes-per-second rate as megabytes/sec with one decimal, e.g. `"3.2"`. */
+export function formatMegabytesPerSecond(bytesPerSecond: number): string {
+  return (bytesPerSecond / (1024 * 1024)).toFixed(1);
+}
diff --git a/services/localAiFacade.ts b/services/localAiFacade.ts
index 89c1f731..96edeea1 100644
--- a/services/localAiFacade.ts
+++ b/services/localAiFacade.ts
@@ -156,7 +156,14 @@ export async function generateLocalText(
         : 0;
       const normalized = { progress: fraction, text: report.text };
       if (canReportToGlobalProgress()) {
-        inferenceProgressEmitter.reportWebLlmProgress(normalized.progress, normalized.text);
+        // QNBS-v3 (#333 item 1): workerModelId is the actual model this attempt is downloading —
+        // captured by closure, safe to read here even though it's assigned after this closure is
+        // defined, since reportProgress is only ever invoked later, once it holds its final value.
+        inferenceProgressEmitter.reportWebLlmProgress(
+          normalized.progress,
+          normalized.text,
+          workerModelId,
+        );
       }
       onProgress?.(normalized);
     };
@@ -395,7 +402,7 @@ export async function preloadLocalModel(
   const startedAt = performance.now();
   try {
     // QNBS-v3: Worker fallback must still make the download visible; otherwise Settings appears frozen.
-    inferenceProgressEmitter.reportWebLlmProgress(0, 'Preparing local model');
+    inferenceProgressEmitter.reportWebLlmProgress(0, 'Preparing local model', modelId);
     const reportProgress = (report: WebLlmProgressReport) => {
       const fraction = Number.isFinite(report.progress)
         ? Math.min(1, Math.max(0, report.progress))
diff --git a/services/voice/voiceCommandService.ts b/services/voice/voiceCommandService.ts
index c2048473..3273a53b 100644
--- a/services/voice/voiceCommandService.ts
+++ b/services/voice/voiceCommandService.ts
@@ -605,10 +605,28 @@ export class VoiceCommandService {
         onProgress: (progress: unknown) => {
           if (signal?.aborted) return;
           const p = progress as { progress?: number; loaded?: number; total?: number };
-          const pct = p.progress ?? (p.loaded && p.total ? p.loaded / p.total : 0);
+          // QNBS-v3 (#333 item 1): transformers.js's own `progress` field is 0-100 (percent), NOT a
+          // 0-1 fraction — verified against its readResponse() source (`progress = loaded/total*100`).
+          // The prior `p.progress ?? ...` fallback used it as-is, so `Math.min(0.95, pct)` clamped to
+          // 0.95 almost immediately (any progress > 0.95%, i.e. after the very first chunk), making
+          // the bar look stuck/frozen for the entire download. Prefer the real loaded/total ratio —
+          // it's also what the new byte-level UI needs — and only fall back to progress/100.
+          const hasBytes =
+            typeof p.loaded === 'number' && typeof p.total === 'number' && p.total > 0;
+          const pct = hasBytes
+            ? (p.loaded as number) / (p.total as number)
+            : typeof p.progress === 'number'
+              ? p.progress / 100
+              : 0;
           this.d(
             settingsActions.setVoiceSettings({
               wasmModelDownloadProgress: Math.min(0.95, pct),
+              ...(hasBytes
+                ? {
+                    wasmModelDownloadLoadedBytes: p.loaded,
+                    wasmModelDownloadTotalBytes: p.total,
+                  }
+                : {}),
             }),
           );
         },
diff --git a/tests/unit/LocalAiDownloadProgress.test.tsx b/tests/unit/LocalAiDownloadProgress.test.tsx
index 9c05f372..955208bc 100644
--- a/tests/unit/LocalAiDownloadProgress.test.tsx
+++ b/tests/unit/LocalAiDownloadProgress.test.tsx
@@ -17,6 +17,10 @@ let mockSnapshot = {
   progress: 0,
   estimatedSecondsRemaining: null as number | null,
   text: '',
+  // QNBS-v3 (#333 item 1)
+  loadedBytes: null as number | null,
+  totalBytes: null as number | null,
+  bytesPerSecond: null as number | null,
 };
 let capturedSubscriber: ((p: typeof mockSnapshot) => void) | null = null;
 
@@ -70,7 +74,15 @@ import { LocalAiDownloadProgress } from '../../components/settings/LocalAiDownlo
 describe('LocalAiDownloadProgress', () => {
   beforeEach(() => {
     vi.clearAllMocks();
-    mockSnapshot = { state: 'idle', progress: 0, estimatedSecondsRemaining: null, text: '' };
+    mockSnapshot = {
+      state: 'idle',
+      progress: 0,
+      estimatedSecondsRemaining: null,
+      text: '',
+      loadedBytes: null,
+      totalBytes: null,
+      bytesPerSecond: null,
+    };
     capturedSubscriber = null;
   });
 
@@ -80,19 +92,43 @@ describe('LocalAiDownloadProgress', () => {
   });
 
   it('renders the progress modal when state is loading', () => {
-    mockSnapshot = { state: 'loading', progress: 0.5, estimatedSecondsRemaining: 30, text: '' };
+    mockSnapshot = {
+      state: 'loading',
+      progress: 0.5,
+      estimatedSecondsRemaining: 30,
+      text: '',
+      loadedBytes: null,
+      totalBytes: null,
+      bytesPerSecond: null,
+    };
     render(<LocalAiDownloadProgress />);
     expect(screen.getByRole('progressbar')).toBeInTheDocument();
   });
 
   it('renders cancel button when loading', () => {
-    mockSnapshot = { state: 'loading', progress: 0.3, estimatedSecondsRemaining: null, text: '' };
+    mockSnapshot = {
+      state: 'loading',
+      progress: 0.3,
+      estimatedSecondsRemaining: null,
+      text: '',
+      loadedBytes: null,
+      totalBytes: null,
+      bytesPerSecond: null,
+    };
     render(<LocalAiDownloadProgress />);
     expect(screen.getByText('settings.ai.localAi.cancelButton')).toBeInTheDocument();
   });
 
   it('shows progress percentage in aria-valuenow', () => {
-    mockSnapshot = { state: 'loading', progress: 0.75, estimatedSecondsRemaining: null, text: '' };
+    mockSnapshot = {
+      state: 'loading',
+      progress: 0.75,
+      estimatedSecondsRemaining: null,
+      text: '',
+      loadedBytes: null,
+      totalBytes: null,
+      bytesPerSecond: null,
+    };
     render(<LocalAiDownloadProgress />);
     const progressbar = screen.getByRole('progressbar');
     expect(progressbar).toHaveAttribute('aria-valuenow', '75');
@@ -104,6 +140,9 @@ describe('LocalAiDownloadProgress', () => {
       progress: 0,
       estimatedSecondsRemaining: null,
       text: 'Download failed',
+      loadedBytes: null,
+      totalBytes: null,
+      bytesPerSecond: null,
     };
     render(<LocalAiDownloadProgress />);
     expect(screen.getByText('Download failed')).toBeInTheDocument();
@@ -112,7 +151,15 @@ describe('LocalAiDownloadProgress', () => {
   it('calls releaseGpu when cancel is clicked', async () => {
     const { gpuResourceManager } = await import('../../services/ai/gpuResourceManager');
     const user = userEvent.setup();
-    mockSnapshot = { state: 'loading', progress: 0.2, estimatedSecondsRemaining: null, text: '' };
+    mockSnapshot = {
+      state: 'loading',
+      progress: 0.2,
+      estimatedSecondsRemaining: null,
+      text: '',
+      loadedBytes: null,
+      totalBytes: null,
+      bytesPerSecond: null,
+    };
     render(<LocalAiDownloadProgress />);
     await user.click(screen.getByText('settings.ai.localAi.cancelButton'));
     expect(gpuResourceManager.releaseGpu).toHaveBeenCalledWith('webllm');
@@ -127,6 +174,9 @@ describe('LocalAiDownloadProgress', () => {
       progress: 0.2,
       estimatedSecondsRemaining: null,
       text: 'Download failed',
+      loadedBytes: null,
+      totalBytes: null,
+      bytesPerSecond: null,
     };
     mockRetryLastPreload.mockResolvedValue(undefined);
     render(<LocalAiDownloadProgress />);
@@ -144,6 +194,9 @@ describe('LocalAiDownloadProgress', () => {
       progress: 0.2,
       estimatedSecondsRemaining: null,
       text: 'Download failed',
+      loadedBytes: null,
+      totalBytes: null,
+      bytesPerSecond: null,
     };
     const retryError = new Error('No local model download is available to retry');
     mockRetryLastPreload.mockRejectedValueOnce(retryError);
@@ -158,7 +211,15 @@ describe('LocalAiDownloadProgress', () => {
   });
 
   it('updates progress when subscriber fires', async () => {
-    mockSnapshot = { state: 'idle', progress: 0, estimatedSecondsRemaining: null, text: '' };
+    mockSnapshot = {
+      state: 'idle',
+      progress: 0,
+      estimatedSecondsRemaining: null,
+      text: '',
+      loadedBytes: null,
+      totalBytes: null,
+      bytesPerSecond: null,
+    };
     render(<LocalAiDownloadProgress />);
     expect(screen.queryByRole('progressbar')).not.toBeInTheDocument();
 
@@ -169,8 +230,45 @@ describe('LocalAiDownloadProgress', () => {
         progress: 0.6,
         estimatedSecondsRemaining: null,
         text: '',
+        loadedBytes: null,
+        totalBytes: null,
+        bytesPerSecond: null,
       });
     }
     await waitFor(() => expect(screen.getByRole('progressbar')).toBeInTheDocument());
   });
+
+  it('shows loaded/total MB and speed once the snapshot has byte metrics', () => {
+    mockSnapshot = {
+      state: 'loading',
+      progress: 0.5,
+      estimatedSecondsRemaining: null,
+      text: '',
+      loadedBytes: 350 * 1024 * 1024,
+      totalBytes: 700 * 1024 * 1024,
+      bytesPerSecond: 3.2 * 1024 * 1024,
+    };
+    render(<LocalAiDownloadProgress />);
+    expect(
+      screen.getByText('settings.ai.localAi.downloadSize:{"loaded":"350","total":"700"}'),
+    ).toBeInTheDocument();
+    expect(
+      screen.getByText('settings.ai.localAi.downloadSpeed:{"speed":"3.2"}'),
+    ).toBeInTheDocument();
+  });
+
+  it('shows no size/speed text when the snapshot has no byte metrics', () => {
+    mockSnapshot = {
+      state: 'loading',
+      progress: 0.5,
+      estimatedSecondsRemaining: null,
+      text: '',
+      loadedBytes: null,
+      totalBytes: null,
+      bytesPerSecond: null,
+    };
+    render(<LocalAiDownloadProgress />);
+    expect(screen.queryByText(/downloadSize/)).not.toBeInTheDocument();
+    expect(screen.queryByText(/downloadSpeed/)).not.toBeInTheDocument();
+  });
 });
diff --git a/tests/unit/components/voice/VoiceModelDownloadModal.test.tsx b/tests/unit/components/voice/VoiceModelDownloadModal.test.tsx
index 0d37dcfc..783c72ef 100644
--- a/tests/unit/components/voice/VoiceModelDownloadModal.test.tsx
+++ b/tests/unit/components/voice/VoiceModelDownloadModal.test.tsx
@@ -3,7 +3,7 @@
  * QNBS-v3: P1 tests for P0-5 voice WASM download UI.
  */
 
-import { fireEvent, render, screen } from '@testing-library/react';
+import { fireEvent, render, screen, waitFor } from '@testing-library/react';
 import type React from 'react';
 import { beforeEach, describe, expect, it, vi } from 'vitest';
 import { VoiceModelDownloadModal } from '../../../../components/voice/VoiceModelDownloadModal';
@@ -16,6 +16,9 @@ vi.mock('../../../../hooks/useTranslation', () => ({
         'voice.modelDownload.title': 'Voice Model Download',
         'voice.modelDownload.description': 'Downloading {{model}} model ({{size}} MB)',
         'voice.modelDownload.progress': '{{percent}}% complete',
+        // QNBS-v3 (#333 item 1):
+        'voice.modelDownload.progressBytes': '{{loaded}} MB of {{total}} MB',
+        'voice.modelDownload.speed': '{{speed}} MB/s',
         'voice.modelDownload.error': 'Download failed: {{error}}',
         'voice.modelDownload.cancel': 'Cancel',
         'voice.modelDownload.retry': 'Retry',
@@ -32,9 +35,25 @@ vi.mock('../../../../hooks/useTranslation', () => ({
   }),
 }));
 
+// QNBS-v3 (#333 item 1): mutable per-test state, selector-aware (not the prior `() => 0` stub) so
+// tests can control progress/loadedBytes/totalBytes independently.
+const mockVoiceState = vi.hoisted(() => ({
+  wasmModelDownloadProgress: 0,
+  wasmModelDownloadLoadedBytes: undefined as number | undefined,
+  wasmModelDownloadTotalBytes: undefined as number | undefined,
+}));
+
 vi.mock('../../../../app/hooks', () => ({
   useAppDispatch: () => vi.fn(),
-  useAppSelector: () => 0,
+  useAppSelector: (selector: (s: unknown) => unknown) =>
+    selector({ settings: { voice: mockVoiceState } }),
+}));
+
+// QNBS-v3 (#333 item 1): never resolves — keeps isDownloading true so the loading-state JSX
+// (progress bar + size/speed text) can be asserted without racing the real download flow, which
+// is already covered directly in tests/unit/services/voice/voiceCommandService.test.ts.
+vi.mock('../../../../services/voice/voiceCommandService', () => ({
+  downloadVoiceModels: vi.fn(() => new Promise(() => {})),
 }));
 
 vi.mock('../../../../features/settings/settingsSlice', () => ({
@@ -102,6 +121,9 @@ describe('VoiceModelDownloadModal', () => {
 
   beforeEach(() => {
     vi.clearAllMocks();
+    mockVoiceState.wasmModelDownloadProgress = 0;
+    mockVoiceState.wasmModelDownloadLoadedBytes = undefined;
+    mockVoiceState.wasmModelDownloadTotalBytes = undefined;
   });
 
   it('renders with correct title', () => {
@@ -149,4 +171,21 @@ describe('VoiceModelDownloadModal', () => {
     render(<VoiceModelDownloadModal {...defaultProps} modelType="tts" />);
     expect(screen.getByText(/Kokoro/)).toBeInTheDocument();
   });
+
+  // QNBS-v3 (#333 item 1)
+  describe('byte-level progress display', () => {
+    it('shows loaded/total MB once both are known', async () => {
+      mockVoiceState.wasmModelDownloadLoadedBytes = 21_000_000;
+      mockVoiceState.wasmModelDownloadTotalBytes = 42_000_000;
+      render(<VoiceModelDownloadModal {...defaultProps} />);
+      await waitFor(() => expect(screen.getByTestId('progress')).toBeInTheDocument());
+      expect(screen.getByText('20 MB of 40 MB')).toBeInTheDocument();
+    });
+
+    it('shows no size text when byte counts are not yet known', async () => {
+      render(<VoiceModelDownloadModal {...defaultProps} />);
+      await waitFor(() => expect(screen.getByTestId('progress')).toBeInTheDocument());
+      expect(screen.queryByText(/MB of/)).not.toBeInTheDocument();
+    });
+  });
 });
diff --git a/tests/unit/localAiFacade.test.ts b/tests/unit/localAiFacade.test.ts
index 7078ce27..0016bf65 100644
--- a/tests/unit/localAiFacade.test.ts
+++ b/tests/unit/localAiFacade.test.ts
@@ -240,7 +240,7 @@ describe('localAiFacade', () => {
       await generateLocalText('prompt', 'm', undefined, undefined, undefined, {
         reportToGlobalProgress: true,
       });
-      expect(progSpy).toHaveBeenCalledWith(0.5, 'half');
+      expect(progSpy).toHaveBeenCalledWith(0.5, 'half', 'm');
       expect(readySpy).toHaveBeenCalled();
       progSpy.mockRestore();
       readySpy.mockRestore();
@@ -357,7 +357,7 @@ describe('localAiFacade', () => {
 
     await expect(preloadLocalModel('Qwen2.5-0.5B')).resolves.toMatchObject({ downloaded: false });
 
-    expect(progressSpy).toHaveBeenCalledWith(0, 'Preparing local model');
+    expect(progressSpy).toHaveBeenCalledWith(0, 'Preparing local model', 'Qwen2.5-0.5B');
     expect(errorSpy).toHaveBeenCalledWith('Local model preload did not complete');
   });
 
diff --git a/tests/unit/services/downloadProgressFormat.test.ts b/tests/unit/services/downloadProgressFormat.test.ts
new file mode 100644
index 00000000..bdc4b083
--- /dev/null
+++ b/tests/unit/services/downloadProgressFormat.test.ts
@@ -0,0 +1,29 @@
+import { describe, expect, it } from 'vitest';
+import {
+  formatMegabytes,
+  formatMegabytesPerSecond,
+} from '../../../services/downloadProgressFormat';
+
+describe('formatMegabytes', () => {
+  it('formats bytes as whole megabytes', () => {
+    expect(formatMegabytes(700 * 1024 * 1024)).toBe('700');
+  });
+
+  it('rounds to the nearest whole megabyte', () => {
+    expect(formatMegabytes(1.6 * 1024 * 1024)).toBe('2');
+  });
+
+  it('handles zero bytes', () => {
+    expect(formatMegabytes(0)).toBe('0');
+  });
+});
+
+describe('formatMegabytesPerSecond', () => {
+  it('formats a byte rate with one decimal', () => {
+    expect(formatMegabytesPerSecond(3.2 * 1024 * 1024)).toBe('3.2');
+  });
+
+  it('handles zero bytes per second', () => {
+    expect(formatMegabytesPerSecond(0)).toBe('0.0');
+  });
+});
diff --git a/tests/unit/services/inferenceProgressEmitter.test.ts b/tests/unit/services/inferenceProgressEmitter.test.ts
index a71a6c50..0bdb4f71 100644
--- a/tests/unit/services/inferenceProgressEmitter.test.ts
+++ b/tests/unit/services/inferenceProgressEmitter.test.ts
@@ -95,4 +95,56 @@ describe('inferenceProgressEmitter', () => {
     const snapshot = inferenceProgressEmitter.getWebLlmLoadingSnapshot();
     expect(snapshot.estimatedSecondsRemaining).toBeNull();
   });
+
+  // QNBS-v3 (#333 item 1): derived byte metrics from a known-model-size lookup table.
+  describe('derived byte metrics', () => {
+    it('computes loadedBytes/totalBytes from a known model id', () => {
+      inferenceProgressEmitter.reportWebLlmProgress(
+        0.5,
+        'Downloading',
+        'Llama-3.2-1B-Instruct-q4f16_1-MLC',
+      );
+      const snapshot = inferenceProgressEmitter.getWebLlmLoadingSnapshot();
+      // 700 MB total (WEBLLM_MODEL_APPROX_MB) at 50% progress.
+      expect(snapshot.totalBytes).toBe(700 * 1024 * 1024);
+      expect(snapshot.loadedBytes).toBe(Math.round(700 * 1024 * 1024 * 0.5));
+    });
+
+    it('leaves byte fields null when no modelId is passed', () => {
+      inferenceProgressEmitter.reportWebLlmProgress(0.5, 'Downloading');
+      const snapshot = inferenceProgressEmitter.getWebLlmLoadingSnapshot();
+      expect(snapshot.loadedBytes).toBeNull();
+      expect(snapshot.totalBytes).toBeNull();
+      expect(snapshot.bytesPerSecond).toBeNull();
+    });
+
+    it('leaves byte fields null for an unrecognized modelId', () => {
+      inferenceProgressEmitter.reportWebLlmProgress(0.5, 'Downloading', 'not-a-real-model');
+      const snapshot = inferenceProgressEmitter.getWebLlmLoadingSnapshot();
+      expect(snapshot.loadedBytes).toBeNull();
+      expect(snapshot.totalBytes).toBeNull();
+    });
+
+    it('clears byte fields on ready/error/reset', () => {
+      inferenceProgressEmitter.reportWebLlmProgress(
+        0.5,
+        'Downloading',
+        'Llama-3.2-1B-Instruct-q4f16_1-MLC',
+      );
+      inferenceProgressEmitter.reportWebLlmReady();
+      let snapshot = inferenceProgressEmitter.getWebLlmLoadingSnapshot();
+      expect(snapshot.loadedBytes).toBeNull();
+      expect(snapshot.totalBytes).toBeNull();
+
+      inferenceProgressEmitter.reportWebLlmProgress(
+        0.5,
+        'Downloading',
+        'Llama-3.2-1B-Instruct-q4f16_1-MLC',
+      );
+      inferenceProgressEmitter.reportWebLlmError('failed');
+      snapshot = inferenceProgressEmitter.getWebLlmLoadingSnapshot();
+      expect(snapshot.loadedBytes).toBeNull();
+      expect(snapshot.totalBytes).toBeNull();
+    });
+  });
 });
diff --git a/tests/unit/services/voice/voiceCommandService.test.ts b/tests/unit/services/voice/voiceCommandService.test.ts
index 734320ac..bd982b7f 100644
--- a/tests/unit/services/voice/voiceCommandService.test.ts
+++ b/tests/unit/services/voice/voiceCommandService.test.ts
@@ -87,6 +87,14 @@ vi.mock('../../../../features/voice/voiceSlice', () => ({
   setWakeWordStatus: vi.fn(),
 }));
 
+// QNBS-v3 (#333 item 1): mock the pipeline factory downloadVoiceModels() calls to trigger the
+// model download — lets tests drive its onProgress callback with representative payloads.
+const { mockPipeline } = vi.hoisted(() => ({ mockPipeline: vi.fn() }));
+vi.mock('@huggingface/transformers', () => ({
+  pipeline: mockPipeline,
+  env: { backends: { onnx: { wasm: {} } } },
+}));
+
 import { VoiceCommandService } from '../../../../services/voice/voiceCommandService';
 
 describe('VoiceCommandService', () => {
@@ -242,4 +250,82 @@ describe('VoiceCommandService', () => {
       expect(disposeSpies.wake).toHaveBeenCalled();
     });
   });
+
+  // QNBS-v3 (#333 item 1): transformers.js's own progress payload is
+  // `{ progress: <0-100>, loaded: <bytes>, total: <bytes> }` (verified against its readResponse()
+  // source — progress = loaded/total*100). The prior code treated `progress` as if it were already
+  // a 0-1 fraction, so `Math.min(0.95, pct)` clamped to 0.95 almost immediately (any progress value
+  // over 0.95%), making the download bar look stuck. These tests pin the fixed 0-1 conversion and
+  // the new real byte-count threading.
+  describe('downloadVoiceModels', () => {
+    let dispatch: ReturnType<typeof vi.fn>;
+
+    beforeEach(() => {
+      dispatch = vi.fn();
+      service.setDispatch(dispatch as never, vi.fn() as never);
+      mockPipeline.mockReset();
+    });
+
+    it('derives a 0-1 progress fraction from loaded/total bytes, not the raw 0-100 percent field', async () => {
+      mockPipeline.mockImplementation(async (_task: string, _model: string, opts: unknown) => {
+        const { onProgress } = opts as { onProgress: (p: unknown) => void };
+        onProgress({ progress: 20, loaded: 8_400_000, total: 42_000_000 });
+        return { dispose: vi.fn() };
+      });
+
+      await service.downloadVoiceModels('stt');
+
+      const progressCalls = dispatch.mock.calls
+        .map((c) => c[0])
+        .filter(
+          (a): a is { payload: { wasmModelDownloadProgress?: number } } =>
+            typeof (a as { payload?: { wasmModelDownloadProgress?: number } })?.payload
+              ?.wasmModelDownloadProgress === 'number',
+        );
+      // 8.4M / 42M = 0.2 — well under the 0.95 clamp, not stuck at 0.95 from a single early tick.
+      expect(progressCalls.some((a) => a.payload.wasmModelDownloadProgress === 0.2)).toBe(true);
+    });
+
+    it('threads real loadedBytes/totalBytes from the progress payload into Redux', async () => {
+      mockPipeline.mockImplementation(async (_task: string, _model: string, opts: unknown) => {
+        const { onProgress } = opts as { onProgress: (p: unknown) => void };
+        onProgress({ progress: 50, loaded: 21_000_000, total: 42_000_000 });
+        return { dispose: vi.fn() };
+      });
+
+      await service.downloadVoiceModels('stt');
+
+      const byteCall = dispatch.mock.calls
+        .map((c) => c[0])
+        .find(
+          (a): a is { payload: { wasmModelDownloadLoadedBytes?: number } } =>
+            typeof (a as { payload?: { wasmModelDownloadLoadedBytes?: number } })?.payload
+              ?.wasmModelDownloadLoadedBytes === 'number',
+        );
+      expect(byteCall?.payload.wasmModelDownloadLoadedBytes).toBe(21_000_000);
+      expect(
+        (byteCall as unknown as { payload: { wasmModelDownloadTotalBytes: number } }).payload
+          .wasmModelDownloadTotalBytes,
+      ).toBe(42_000_000);
+    });
+
+    it('falls back to progress/100 when the payload has no loaded/total (defensive, not the normal path)', async () => {
+      mockPipeline.mockImplementation(async (_task: string, _model: string, opts: unknown) => {
+        const { onProgress } = opts as { onProgress: (p: unknown) => void };
+        onProgress({ progress: 30 });
+        return { dispose: vi.fn() };
+      });
+
+      await service.downloadVoiceModels('stt');
+
+      const progressCalls = dispatch.mock.calls
+        .map((c) => c[0])
+        .filter(
+          (a): a is { payload: { wasmModelDownloadProgress?: number } } =>
+            typeof (a as { payload?: { wasmModelDownloadProgress?: number } })?.payload
+              ?.wasmModelDownloadProgress === 'number',
+        );
+      expect(progressCalls.some((a) => a.payload.wasmModelDownloadProgress === 0.3)).toBe(true);
+    });
+  });
 });
diff --git a/types.ts b/types.ts
index e76e2359..b9bc5f7d 100644
--- a/types.ts
+++ b/types.ts
@@ -609,6 +609,12 @@ export interface VoiceSettings {
   webSpeechConsentGranted?: boolean;
   /** WASM voice models download progress (0-1) */
   wasmModelDownloadProgress?: number;
+  /** QNBS-v3 (#333 item 1): real bytes loaded so far, from transformers.js's own progress payload
+   * (not an approximation, unlike the WebLLM text-model path — this download is a raw fetch the
+   * library already reports byte counts for). Undefined until the first progress tick arrives. */
+  wasmModelDownloadLoadedBytes?: number;
+  /** QNBS-v3 (#333 item 1): real total byte size of the file currently downloading. */
+  wasmModelDownloadTotalBytes?: number;
   /** WASM voice models ready for use */
   wasmModelsReady?: boolean;
   /** WASM voice model download error message */

From 54bd0dca7f445fe22102c6b881ff2053bb3b6903 Mon Sep 17 00:00:00 2001
From: qnbs <155236708+qnbs@users.noreply.github.com>
Date: Wed, 12 Aug 2026 22:02:18 +0200
Subject: [PATCH 2/3] chore(security): ignore extract-zip GHSA-jmr9-qjv8-65gv
 (no fix available)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Freshly published/reviewed 2026-08-12; latest extract-zip release (2.0.1) has
no patched version to override to. Transitive devDependency of
@puppeteer/browsers (Playwright's browser-binary downloader) — only ever
extracts Playwright/Chromium's own CDN-hosted zip releases, never a user- or
attacker-supplied archive, and ships in no production bundle. Documented in
src-tauri/osv-scanner.toml's IgnoredVulns list (existing pattern for
unfixable transitive findings) and AUDIT.md.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
---
 AUDIT.md                   |  1 +
 src-tauri/osv-scanner.toml | 27 +++++++++++++++++++++------
 2 files changed, 22 insertions(+), 6 deletions(-)

diff --git a/AUDIT.md b/AUDIT.md
index da15f93f..133eb80b 100644
--- a/AUDIT.md
+++ b/AUDIT.md
@@ -98,6 +98,7 @@ the same root cause in one sprint is itself the finding: **the failure mode is n
 ## v1.24.0 Dependency Hygiene + Onboarding + Docs Truth-up (2026-06-21, PR F)
 
 - **`joi` override (accepted risk, KEPT):** `pnpm-workspace.yaml` `overrides.joi: ^18.2.1` pins the patched `joi` pulled transitively via `wait-on` (Storybook/test-runner wait helper), mitigating **GHSA-q7cg-457f-vx79** (unpublished jsdom exposure in `@hapi/statehood`). Still required — `wait-on@9.x` still depends on `joi`. `pnpm audit --audit-level=high` clean with the override in place; rationale documented inline in `pnpm-workspace.yaml` and here.
+- **`extract-zip@2.0.1` OSV ignore (accepted risk, 2026-08-12):** **GHSA-jmr9-qjv8-65gv** / CVE-2026-56876 (CVSS 8.6, unvalidated symlink path traversal when extracting an attacker-controlled zip) was published/GitHub-reviewed 2026-08-12, freshly flagging a transitive devDependency of `@puppeteer/browsers` (Playwright's browser-binary downloader). No fixed version exists (`extract-zip@2.0.1` is the final release — `pnpm.overrides` cannot remediate an unpatched advisory), so `pnpm.overrides` doesn't apply here; documented as an `IgnoredVulns` entry in `src-tauri/osv-scanner.toml` instead, matching the file's existing pattern for unfixable transitive findings. Not exploitable in this project: only ever extracts Playwright/Chromium's own CDN-hosted zip releases, never a user- or attacker-supplied archive, and ships in no production bundle.
 - **SBOM — deferred (decision):** evaluated a `@cyclonedx/cyclonedx-npm` generate-on-tag step; **not adopted in 1.24.0** to keep the release scope tight. Socket Security (PR + project report) already runs every CI run and provides dependency-risk + an SBOM dashboard, so the marginal value is low. Revisit when a formal SBOM artifact is required by a downstream consumer.
 - **README metric drift fixed:** `scripts/sync-readme-metrics.mjs` had its locale count **hard-coded to `11`**, so its regexes stopped matching after the 11→17 expansion and silently froze the key count at a stale value. Locale count is now **dynamic** (counts `locales/` dirs) and the regexes match any digit count; re-run → README reads **2786 keys × 17 locales** with the drift guard green.
 - **Docs truth-up:** corrected the stale `public/sw.js` "must hand-sync `APP_VERSION`" note in `CLAUDE.md` (it is auto-synced by `scripts/sync-sw-version.mjs` + `sync-tauri-version.mjs` via `predev`/`prebuild`); refreshed the stale 5-locale / `2 594 keys × 11 locales` strings in `CONTRIBUTING.md` + `.github/copilot-instructions.md` to 17 locales.
diff --git a/src-tauri/osv-scanner.toml b/src-tauri/osv-scanner.toml
index 63cf709c..5ed060f0 100644
--- a/src-tauri/osv-scanner.toml
+++ b/src-tauri/osv-scanner.toml
@@ -117,9 +117,24 @@ ignoreUntil = "2026-11-30T00:00:00Z"
 reason = "unic-ucd-ident 0.9.0: unmaintained Unicode data crate; final version; no exploitable attack surface"
 
 # ─── npm advisories (pnpm-lock.yaml) ─────────────────────────────────────────
-# Reviewed 2026-06-16 / 2026-07-26. All currently-flagged npm advisories are
-# remediated via pnpm-workspace.yaml overrides (tmp, ws, form-data, protobufjs,
-# @babel/core, dompurify→3.4.12, js-yaml→4.x, body-parser→1.20.6, …) — no open
-# npm ignores remain. (The former dompurify GHSA-x4vx-rjvf-j5p4 and js-yaml-3.x
-# GHSA-h67p-54hq-rp68 ignores were dropped: both advisories no longer match the
-# lockfile after the overrides.)
+# Reviewed 2026-06-16 / 2026-07-26 / 2026-08-12. Most currently-flagged npm
+# advisories are remediated via pnpm-workspace.yaml overrides (tmp, ws,
+# form-data, protobufjs, @babel/core, dompurify→3.4.12, js-yaml→4.x,
+# body-parser→1.20.6, …). (The former dompurify GHSA-x4vx-rjvf-j5p4 and
+# js-yaml-3.x GHSA-h67p-54hq-rp68 ignores were dropped: both advisories no
+# longer match the lockfile after the overrides.) One documented ignore
+# remains below — no fixed version exists to override to.
+
+# extract-zip@2.0.1 (CVE-2026-56876 / GHSA-jmr9-qjv8-65gv, CVSS 8.6): unvalidated
+# symlink path traversal when extracting an attacker-controlled zip archive.
+# extract-zip@2.0.1 is the latest published release — no patched version exists
+# (first_patched_version: null on the advisory) — so pnpm.overrides cannot fix
+# this. It's a transitive devDependency of @puppeteer/browsers (Playwright's
+# browser-binary downloader): only ever extracts Playwright/Chromium's own
+# CDN-hosted zip releases, never a user- or attacker-supplied archive, and never
+# ships in the production app bundle. No exploitable attack surface here.
+
+[[IgnoredVulns]]
+id = "GHSA-jmr9-qjv8-65gv"
+ignoreUntil = "2026-11-30T00:00:00Z"
+reason = "extract-zip 2.0.1: unvalidated symlink path traversal; latest release, no fixed version exists; transitive Playwright/@puppeteer/browsers devDependency only extracting Playwright's own CDN-hosted browser binaries, never user-supplied zips; no runtime footprint in the shipped app"

From c640b70cc46ba60f9071d72630e4b926ce5a2141 Mon Sep 17 00:00:00 2001
From: qnbs <155236708+qnbs@users.noreply.github.com>
Date: Thu, 13 Aug 2026 03:42:41 +0200
Subject: [PATCH 3/3] fix: PR review correction loop for #346 (download
 progress metrics)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Addresses all 14 CodeRabbit/CodeAnt/Qodo/Sourcery review findings on PR #346:

- services/voice/voiceCommandService.ts + components/voice/VoiceModelDownloadModal.tsx:
  explicitly clear wasmModelDownloadLoadedBytes/wasmModelDownloadTotalBytes at every
  download lifecycle boundary (start, cancel, error, completion) — setVoiceSettings
  merges partial state, so a prior attempt's bytes previously lingered in Redux until
  the next byte-bearing progress tick arrived, showing stale size/speed on retry or
  when switching models. types.ts: widened both fields to `number | undefined` so
  they can be explicitly cleared under exactOptionalPropertyTypes.
- components/voice/VoiceModelDownloadModal.tsx: bytesPerSecond is now delta-based
  (bytes and timestamp of the previous sample, reset when downloading stops) instead
  of cumulative-loaded-bytes / elapsed-since-start, which included startup time and
  smoothed real throughput changes into a misleading average presented as a live rate.
- services/downloadProgressFormat.ts: formatMegabytesPerSecond renamed to
  megabytesPerSecond and now returns a number instead of a toFixed() string — the
  decimal separator is locale-dependent (comma vs period), so callers now format it
  via formatNumber() from useTranslation() instead of a hardcoded period.
- packages/ai-core/src/index.ts / services/ai/localModelStorageService.ts: removed
  the duplicate WEBLLM_MODEL_APPROX_MB literal from localModelStorageService — it now
  re-exports the canonical ai-core table instead of maintaining a second, driftable
  copy of the same model-id-to-size mapping.
- services/ai/inferenceProgressEmitter.ts: replaced an unsafe `as WebLlmModelId` cast
  with a proper type guard (isKnownWebLlmModel) — modelId stays `string` since real
  call sites in localAiFacade.ts pass broader adaptive/fallback model ids, not just
  the curated WebLlmModelId union, so narrowing the param type would have broken them.
- components/settings/LocalAiDownloadProgress.tsx: prefixes WebLLM's derived/estimated
  byte and speed values with "~" so users can distinguish them from the voice download
  UI's measured bytes, without adding new i18n keys (reuses the tilde convention
  already used in WEBLLM_SUPPORTED_MODELS' own size labels).
- 8 locales (el, eu, fa, fi, pt, ru, sv, zh): translated the "of" connector in
  settings.ai.localAi.downloadSize / voice.modelDownload.progressBytes, mirroring
  this file's own already-translated storageUsage key's per-locale phrasing exactly.
  MB/s stays untranslated — a universal unit abbreviation, same as every other locale
  including production ones (de/es/fr/it/ar/he/...) already leave it.
- New/updated tests: voiceCommandService (3 new lifecycle-boundary byte-clearing
  cases), VoiceModelDownloadModal (delta-speed test, byte-clear-on-cancel coverage),
  downloadProgressFormat (renamed function, plain-number-return assertion).

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
---
 .../settings/LocalAiDownloadProgress.tsx      | 12 ++--
 components/voice/VoiceModelDownloadModal.tsx  | 46 ++++++++----
 locales/el/settings.json                      |  4 +-
 locales/eu/settings.json                      |  4 +-
 locales/fa/settings.json                      |  4 +-
 locales/fi/settings.json                      |  4 +-
 locales/pt/settings.json                      |  4 +-
 locales/ru/settings.json                      |  4 +-
 locales/sv/settings.json                      |  4 +-
 locales/zh/settings.json                      |  4 +-
 public/locales/el/bundle.json                 |  4 +-
 public/locales/eu/bundle.json                 |  4 +-
 public/locales/fa/bundle.json                 |  4 +-
 public/locales/fi/bundle.json                 |  4 +-
 public/locales/pt/bundle.json                 |  4 +-
 public/locales/ru/bundle.json                 |  4 +-
 public/locales/sv/bundle.json                 |  4 +-
 public/locales/zh/bundle.json                 |  4 +-
 services/ai/inferenceProgressEmitter.ts       | 10 ++-
 services/ai/localModelStorageService.ts       | 19 ++---
 services/downloadProgressFormat.ts            |  9 ++-
 services/voice/voiceCommandService.ts         | 11 ++-
 tests/unit/LocalAiDownloadProgress.test.tsx   |  6 +-
 .../voice/VoiceModelDownloadModal.test.tsx    | 31 +++++++-
 .../services/downloadProgressFormat.test.ts   | 18 ++---
 .../voice/voiceCommandService.test.ts         | 72 +++++++++++++++++++
 types.ts                                      |  7 +-
 27 files changed, 223 insertions(+), 82 deletions(-)

diff --git a/components/settings/LocalAiDownloadProgress.tsx b/components/settings/LocalAiDownloadProgress.tsx
index be941428..bded5029 100644
--- a/components/settings/LocalAiDownloadProgress.tsx
+++ b/components/settings/LocalAiDownloadProgress.tsx
@@ -9,11 +9,11 @@ import {
   inferenceProgressEmitter,
   type WebLlmLoadProgress,
 } from '../../services/ai/inferenceProgressEmitter';
-import { formatMegabytes, formatMegabytesPerSecond } from '../../services/downloadProgressFormat';
+import { formatMegabytes, megabytesPerSecond } from '../../services/downloadProgressFormat';
 import { abortActivePreload, retryLastPreload } from '../../services/localAiFacade';
 
 export const LocalAiDownloadProgress: FC = () => {
-  const { t } = useTranslation();
+  const { t, formatNumber } = useTranslation();
   const titleId = useId();
   const cancelRef = useRef<HTMLButtonElement>(null);
   const [progress, setProgress] = useState<WebLlmLoadProgress>(
@@ -49,17 +49,19 @@ export const LocalAiDownloadProgress: FC = () => {
 
   // QNBS-v3 (#333 item 1): approximate — see WebLlmLoadProgress's own doc comment. Only shown once
   // both bounds are known (null whenever the active model id isn't in WEBLLM_MODEL_APPROX_MB).
+  // QNBS-v3 (#333/Qodo): "~" prefix marks these as derived estimates, unlike the voice download UI's
+  // measured bytes — avoids a new i18n key by reusing the tilde convention already in model labels.
   const sizeText =
     progress.loadedBytes != null && progress.totalBytes != null
       ? t<string>('settings.ai.localAi.downloadSize', {
-          loaded: formatMegabytes(progress.loadedBytes),
-          total: formatMegabytes(progress.totalBytes),
+          loaded: `~${formatMegabytes(progress.loadedBytes)}`,
+          total: `~${formatMegabytes(progress.totalBytes)}`,
         })
       : '';
   const speedText =
     progress.bytesPerSecond != null
       ? t<string>('settings.ai.localAi.downloadSpeed', {
-          speed: formatMegabytesPerSecond(progress.bytesPerSecond),
+          speed: `~${formatNumber(megabytesPerSecond(progress.bytesPerSecond), { minimumFractionDigits: 1, maximumFractionDigits: 1 })}`,
         })
       : '';
 
diff --git a/components/voice/VoiceModelDownloadModal.tsx b/components/voice/VoiceModelDownloadModal.tsx
index 2d9e069c..6841c2ae 100644
--- a/components/voice/VoiceModelDownloadModal.tsx
+++ b/components/voice/VoiceModelDownloadModal.tsx
@@ -7,7 +7,7 @@ import React, { useCallback, useEffect, useRef, useState } from 'react';
 import { useAppDispatch, useAppSelector } from '../../app/hooks';
 import { settingsActions } from '../../features/settings/settingsSlice';
 import { useTranslation } from '../../hooks/useTranslation';
-import { formatMegabytes, formatMegabytesPerSecond } from '../../services/downloadProgressFormat';
+import { formatMegabytes, megabytesPerSecond } from '../../services/downloadProgressFormat';
 import { Button } from '../ui/Button';
 import { Modal } from '../ui/Modal';
 import { Progress } from '../ui/Progress';
@@ -29,7 +29,7 @@ export const VoiceModelDownloadModal = React.memo(function VoiceModelDownloadMod
   onClose,
   modelType,
 }: VoiceModelDownloadModalProps) {
-  const { t } = useTranslation();
+  const { t, formatNumber } = useTranslation();
   const dispatch = useAppDispatch();
   const progress = useAppSelector((s) => s.settings.voice.wasmModelDownloadProgress ?? 0);
   // QNBS-v3 (#333 item 1): real bytes from transformers.js's own progress payload (not an
@@ -40,23 +40,29 @@ export const VoiceModelDownloadModal = React.memo(function VoiceModelDownloadMod
   const [error, setError] = useState<string | null>(null);
   // QNBS-v3: CodeAnt P2-3 — AbortController lets cancel button stop an in-flight download
   const abortRef = useRef<AbortController | null>(null);
-  // QNBS-v3 (#333 item 1): wall-clock speed, computed client-side from real loaded-bytes deltas —
-  // the service layer only reports cumulative bytes, not a rate.
-  const downloadStartMsRef = useRef<number | null>(null);
+  // QNBS-v3 (#333/CodeRabbit+Qodo): delta-based rate from consecutive byte samples — the prior
+  // cumulative-bytes/elapsed-since-start calc included startup time and smoothed out real throughput
+  // changes into a misleading average, not a live rate.
+  const prevSampleRef = useRef<{ bytes: number; atMs: number } | null>(null);
   const [bytesPerSecond, setBytesPerSecond] = useState<number | null>(null);
 
   useEffect(() => {
     if (!isDownloading) {
-      downloadStartMsRef.current = null;
+      prevSampleRef.current = null;
       setBytesPerSecond(null);
       return;
     }
-    if (downloadStartMsRef.current == null) {
-      downloadStartMsRef.current = Date.now();
-    }
     if (loadedBytes == null) return;
-    const elapsedSeconds = (Date.now() - downloadStartMsRef.current) / 1000;
-    setBytesPerSecond(elapsedSeconds > 0.5 ? loadedBytes / elapsedSeconds : null);
+    const now = Date.now();
+    const prev = prevSampleRef.current;
+    if (prev) {
+      const elapsedSeconds = (now - prev.atMs) / 1000;
+      const deltaBytes = loadedBytes - prev.bytes;
+      if (elapsedSeconds > 0.5 && deltaBytes >= 0) {
+        setBytesPerSecond(deltaBytes / elapsedSeconds);
+      }
+    }
+    prevSampleRef.current = { bytes: loadedBytes, atMs: now };
   }, [isDownloading, loadedBytes]);
 
   const handleDownload = useCallback(async () => {
@@ -94,8 +100,15 @@ export const VoiceModelDownloadModal = React.memo(function VoiceModelDownloadMod
     abortRef.current?.abort();
     abortRef.current = null;
     setIsDownloading(false);
-    // QNBS-v3: CodeAnt — reset progress so a reopened modal auto-starts (it only fires at progress 0).
-    dispatch(settingsActions.setVoiceSettings({ wasmModelDownloadProgress: 0 }));
+    // QNBS-v3 (#333/CodeAnt+Qodo): reset progress AND byte counts — the service's own abort branch
+    // doesn't dispatch, so a cancel mid-download would otherwise leave stale bytes for the next attempt.
+    dispatch(
+      settingsActions.setVoiceSettings({
+        wasmModelDownloadProgress: 0,
+        wasmModelDownloadLoadedBytes: undefined,
+        wasmModelDownloadTotalBytes: undefined,
+      }),
+    );
     onClose();
   }, [dispatch, onClose]);
 
@@ -121,7 +134,12 @@ export const VoiceModelDownloadModal = React.memo(function VoiceModelDownloadMod
       : '';
   const speedText =
     bytesPerSecond != null
-      ? t('voice.modelDownload.speed', { speed: formatMegabytesPerSecond(bytesPerSecond) })
+      ? t('voice.modelDownload.speed', {
+          speed: formatNumber(megabytesPerSecond(bytesPerSecond), {
+            minimumFractionDigits: 1,
+            maximumFractionDigits: 1,
+          }),
+        })
       : '';
 
   return (
diff --git a/locales/el/settings.json b/locales/el/settings.json
index ade079b0..d9bf1ee4 100644
--- a/locales/el/settings.json
+++ b/locales/el/settings.json
@@ -159,7 +159,7 @@
   "settings.ai.localAi.downloadCancelled": "Η λήψη ακυρώθηκε",
   "settings.ai.localAi.downloadEta": "περίπου. Απομένουν {{seconds}}s",
   "settings.ai.localAi.downloadFailed": "Η λήψη απέτυχε. Δοκιμάστε ξανά.",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB από {{total}} MB",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Λήψη μοντέλου AI…",
   "settings.ai.localAi.downloadedLabel": "κατεβάσει",
@@ -821,7 +821,7 @@
   "voice.modelDownload.description": "Λήψη μοντέλου {{model}} ({{size}} MB). Αυτό επιτρέπει την τοπική επεξεργασία φωνής χωρίς δρομολόγηση cloud.",
   "voice.modelDownload.error": "Η λήψη απέτυχε: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% ολοκληρώθηκε",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB από {{total}} MB",
   "voice.modelDownload.retry": "Δοκιμάζω πάλι",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Λήψη φωνητικού μοντέλου"
diff --git a/locales/eu/settings.json b/locales/eu/settings.json
index da932878..41a1e720 100644
--- a/locales/eu/settings.json
+++ b/locales/eu/settings.json
@@ -159,7 +159,7 @@
   "settings.ai.localAi.downloadCancelled": "Deskarga bertan behera utzi da",
   "settings.ai.localAi.downloadEta": "gutxi gorabehera. {{seconds}}s geratzen dira",
   "settings.ai.localAi.downloadFailed": "Deskargatu huts egin du. Mesedez, saiatu berriro.",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB/{{total}} MB",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "AI eredua deskargatzen…",
   "settings.ai.localAi.downloadedLabel": "deskargatu",
@@ -821,7 +821,7 @@
   "voice.modelDownload.description": "{{model}} eredua deskargatzen ({{size}} MB). Honek ahots lokala prozesatzea ahalbidetzen du hodeiko bideratzerik gabe.",
   "voice.modelDownload.error": "Deskargatu huts egin du: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% osatuta",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB/{{total}} MB",
   "voice.modelDownload.retry": "Saiatu berriro",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Ahots eredua deskargatu"
diff --git a/locales/fa/settings.json b/locales/fa/settings.json
index 6f87bc8d..8316d83e 100644
--- a/locales/fa/settings.json
+++ b/locales/fa/settings.json
@@ -159,7 +159,7 @@
   "settings.ai.localAi.downloadCancelled": "دانلود لغو شد",
   "settings.ai.localAi.downloadEta": "تقریبا {{seconds}} باقی مانده است",
   "settings.ai.localAi.downloadFailed": "دانلود انجام نشد. لطفا دوباره امتحان کنید.",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} مگابایت از {{total}} مگابایت",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "در حال دانلود مدل هوش مصنوعی…",
   "settings.ai.localAi.downloadedLabel": "دانلود شده است",
@@ -821,7 +821,7 @@
   "voice.modelDownload.description": "در حال دانلود مدل {{model}} ({{size}} مگابایت). این امکان پردازش صدای محلی را بدون مسیریابی ابری فراهم می کند.",
   "voice.modelDownload.error": "دانلود انجام نشد: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% کامل شده است",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} مگابایت از {{total}} مگابایت",
   "voice.modelDownload.retry": "دوباره امتحان کنید",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "دانلود مدل صوتی"
diff --git a/locales/fi/settings.json b/locales/fi/settings.json
index c0e6813e..782cb60d 100644
--- a/locales/fi/settings.json
+++ b/locales/fi/settings.json
@@ -159,7 +159,7 @@
   "settings.ai.localAi.downloadCancelled": "Lataus peruutettu",
   "settings.ai.localAi.downloadEta": "noin {{seconds}}s jäljellä",
   "settings.ai.localAi.downloadFailed": "Lataus epäonnistui. Yritä uudelleen.",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} Mt / {{total}} Mt",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Ladataan tekoälymallia…",
   "settings.ai.localAi.downloadedLabel": "ladattu",
@@ -821,7 +821,7 @@
   "voice.modelDownload.description": "Ladataan mallia {{model}} ({{size}} Mt). Tämä mahdollistaa paikallisen äänenkäsittelyn ilman pilvireititystä.",
   "voice.modelDownload.error": "Lataus epäonnistui: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% valmiina",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} Mt / {{total}} Mt",
   "voice.modelDownload.retry": "Yritä uudelleen",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Äänimallin lataus"
diff --git a/locales/pt/settings.json b/locales/pt/settings.json
index 179c9291..55fdb84d 100644
--- a/locales/pt/settings.json
+++ b/locales/pt/settings.json
@@ -159,7 +159,7 @@
   "settings.ai.localAi.downloadCancelled": "Download cancelado",
   "settings.ai.localAi.downloadEta": "aprox. {{seconds}} restantes",
   "settings.ai.localAi.downloadFailed": "A transferência falhou. Tente novamente.",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB de {{total}} MB",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Downloading IA model…",
   "settings.ai.localAi.downloadedLabel": "baixado",
@@ -821,7 +821,7 @@
   "voice.modelDownload.description": "Fazendo download do modelo {{model}} ({{size}} MB). Isso permite o processamento de voz local sem roteamento na nuvem.",
   "voice.modelDownload.error": "Falha no download: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% concluído",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB de {{total}} MB",
   "voice.modelDownload.retry": "Tentar novamente",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Download do modelo de voz"
diff --git a/locales/ru/settings.json b/locales/ru/settings.json
index 6862669c..ae23a440 100644
--- a/locales/ru/settings.json
+++ b/locales/ru/settings.json
@@ -159,7 +159,7 @@
   "settings.ai.localAi.downloadCancelled": "Загрузка отменена",
   "settings.ai.localAi.downloadEta": "ок. осталось {{seconds}}с",
   "settings.ai.localAi.downloadFailed": "Загрузка не удалась. Пожалуйста, попробуйте еще раз.",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} МБ из {{total}} МБ",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Загрузка модели ИИ…",
   "settings.ai.localAi.downloadedLabel": "скачал",
@@ -821,7 +821,7 @@
   "voice.modelDownload.description": "Скачиваем модель {{model}} ({{size}} МБ). Это обеспечивает локальную обработку голоса без облачной маршрутизации.",
   "voice.modelDownload.error": "Загрузка не удалась: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% выполнено",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} МБ из {{total}} МБ",
   "voice.modelDownload.retry": "Повторить попытку",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Скачать модель голоса"
diff --git a/locales/sv/settings.json b/locales/sv/settings.json
index da65676d..fc79b9f6 100644
--- a/locales/sv/settings.json
+++ b/locales/sv/settings.json
@@ -159,7 +159,7 @@
   "settings.ai.localAi.downloadCancelled": "Nedladdningen avbröts",
   "settings.ai.localAi.downloadEta": "ca. {{seconds}}s kvar",
   "settings.ai.localAi.downloadFailed": "Nedladdningen misslyckades. Försök igen.",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB av {{total}} MB",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Laddar ner AI-modell...",
   "settings.ai.localAi.downloadedLabel": "nedladdade",
@@ -821,7 +821,7 @@
   "voice.modelDownload.description": "Laddar ner {{model}}-modellen ({{size}} MB). Detta möjliggör lokal röstbehandling utan molnrouting.",
   "voice.modelDownload.error": "Nedladdning misslyckades: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% klar",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB av {{total}} MB",
   "voice.modelDownload.retry": "Försöka igen",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Ladda ner röstmodell"
diff --git a/locales/zh/settings.json b/locales/zh/settings.json
index 52d9500b..fe16dd47 100644
--- a/locales/zh/settings.json
+++ b/locales/zh/settings.json
@@ -159,7 +159,7 @@
   "settings.ai.localAi.downloadCancelled": "下载已取消",
   "settings.ai.localAi.downloadEta": "约。还剩 {{seconds}} 秒",
   "settings.ai.localAi.downloadFailed": "下载失败。请重试。",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB／{{total}} MB",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "正在下载 AI 模型...",
   "settings.ai.localAi.downloadedLabel": "下载的",
@@ -821,7 +821,7 @@
   "voice.modelDownload.description": "正在下载 {{model}} 模型 ({{size}} MB)。这使得本地语音处理无需云路由。",
   "voice.modelDownload.error": "下载失败：{{error}}",
   "voice.modelDownload.progress": "{{percent}}% 完成",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB／{{total}} MB",
   "voice.modelDownload.retry": "重试",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "语音模型下载"
diff --git a/public/locales/el/bundle.json b/public/locales/el/bundle.json
index 3aa160b3..c9bdc928 100644
--- a/public/locales/el/bundle.json
+++ b/public/locales/el/bundle.json
@@ -1805,7 +1805,7 @@
   "settings.ai.localAi.downloadCancelled": "Η λήψη ακυρώθηκε",
   "settings.ai.localAi.downloadEta": "περίπου. Απομένουν {{seconds}}s",
   "settings.ai.localAi.downloadFailed": "Η λήψη απέτυχε. Δοκιμάστε ξανά.",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB από {{total}} MB",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Λήψη μοντέλου AI…",
   "settings.ai.localAi.downloadedLabel": "κατεβάσει",
@@ -2467,7 +2467,7 @@
   "voice.modelDownload.description": "Λήψη μοντέλου {{model}} ({{size}} MB). Αυτό επιτρέπει την τοπική επεξεργασία φωνής χωρίς δρομολόγηση cloud.",
   "voice.modelDownload.error": "Η λήψη απέτυχε: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% ολοκληρώθηκε",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB από {{total}} MB",
   "voice.modelDownload.retry": "Δοκιμάζω πάλι",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Λήψη φωνητικού μοντέλου",
diff --git a/public/locales/eu/bundle.json b/public/locales/eu/bundle.json
index d11d19a4..54513dbf 100644
--- a/public/locales/eu/bundle.json
+++ b/public/locales/eu/bundle.json
@@ -1805,7 +1805,7 @@
   "settings.ai.localAi.downloadCancelled": "Deskarga bertan behera utzi da",
   "settings.ai.localAi.downloadEta": "gutxi gorabehera. {{seconds}}s geratzen dira",
   "settings.ai.localAi.downloadFailed": "Deskargatu huts egin du. Mesedez, saiatu berriro.",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB/{{total}} MB",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "AI eredua deskargatzen…",
   "settings.ai.localAi.downloadedLabel": "deskargatu",
@@ -2467,7 +2467,7 @@
   "voice.modelDownload.description": "{{model}} eredua deskargatzen ({{size}} MB). Honek ahots lokala prozesatzea ahalbidetzen du hodeiko bideratzerik gabe.",
   "voice.modelDownload.error": "Deskargatu huts egin du: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% osatuta",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB/{{total}} MB",
   "voice.modelDownload.retry": "Saiatu berriro",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Ahots eredua deskargatu",
diff --git a/public/locales/fa/bundle.json b/public/locales/fa/bundle.json
index e15bd4c9..bf51af43 100644
--- a/public/locales/fa/bundle.json
+++ b/public/locales/fa/bundle.json
@@ -1805,7 +1805,7 @@
   "settings.ai.localAi.downloadCancelled": "دانلود لغو شد",
   "settings.ai.localAi.downloadEta": "تقریبا {{seconds}} باقی مانده است",
   "settings.ai.localAi.downloadFailed": "دانلود انجام نشد. لطفا دوباره امتحان کنید.",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} مگابایت از {{total}} مگابایت",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "در حال دانلود مدل هوش مصنوعی…",
   "settings.ai.localAi.downloadedLabel": "دانلود شده است",
@@ -2467,7 +2467,7 @@
   "voice.modelDownload.description": "در حال دانلود مدل {{model}} ({{size}} مگابایت). این امکان پردازش صدای محلی را بدون مسیریابی ابری فراهم می کند.",
   "voice.modelDownload.error": "دانلود انجام نشد: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% کامل شده است",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} مگابایت از {{total}} مگابایت",
   "voice.modelDownload.retry": "دوباره امتحان کنید",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "دانلود مدل صوتی",
diff --git a/public/locales/fi/bundle.json b/public/locales/fi/bundle.json
index b2889564..bf29e8c6 100644
--- a/public/locales/fi/bundle.json
+++ b/public/locales/fi/bundle.json
@@ -1805,7 +1805,7 @@
   "settings.ai.localAi.downloadCancelled": "Lataus peruutettu",
   "settings.ai.localAi.downloadEta": "noin {{seconds}}s jäljellä",
   "settings.ai.localAi.downloadFailed": "Lataus epäonnistui. Yritä uudelleen.",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} Mt / {{total}} Mt",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Ladataan tekoälymallia…",
   "settings.ai.localAi.downloadedLabel": "ladattu",
@@ -2467,7 +2467,7 @@
   "voice.modelDownload.description": "Ladataan mallia {{model}} ({{size}} Mt). Tämä mahdollistaa paikallisen äänenkäsittelyn ilman pilvireititystä.",
   "voice.modelDownload.error": "Lataus epäonnistui: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% valmiina",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} Mt / {{total}} Mt",
   "voice.modelDownload.retry": "Yritä uudelleen",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Äänimallin lataus",
diff --git a/public/locales/pt/bundle.json b/public/locales/pt/bundle.json
index 428b7816..0a5fc526 100644
--- a/public/locales/pt/bundle.json
+++ b/public/locales/pt/bundle.json
@@ -1805,7 +1805,7 @@
   "settings.ai.localAi.downloadCancelled": "Download cancelado",
   "settings.ai.localAi.downloadEta": "aprox. {{seconds}} restantes",
   "settings.ai.localAi.downloadFailed": "A transferência falhou. Tente novamente.",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB de {{total}} MB",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Downloading IA model…",
   "settings.ai.localAi.downloadedLabel": "baixado",
@@ -2467,7 +2467,7 @@
   "voice.modelDownload.description": "Fazendo download do modelo {{model}} ({{size}} MB). Isso permite o processamento de voz local sem roteamento na nuvem.",
   "voice.modelDownload.error": "Falha no download: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% concluído",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB de {{total}} MB",
   "voice.modelDownload.retry": "Tentar novamente",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Download do modelo de voz",
diff --git a/public/locales/ru/bundle.json b/public/locales/ru/bundle.json
index feacb2da..ff2ff6d2 100644
--- a/public/locales/ru/bundle.json
+++ b/public/locales/ru/bundle.json
@@ -1805,7 +1805,7 @@
   "settings.ai.localAi.downloadCancelled": "Загрузка отменена",
   "settings.ai.localAi.downloadEta": "ок. осталось {{seconds}}с",
   "settings.ai.localAi.downloadFailed": "Загрузка не удалась. Пожалуйста, попробуйте еще раз.",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} МБ из {{total}} МБ",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Загрузка модели ИИ…",
   "settings.ai.localAi.downloadedLabel": "скачал",
@@ -2467,7 +2467,7 @@
   "voice.modelDownload.description": "Скачиваем модель {{model}} ({{size}} МБ). Это обеспечивает локальную обработку голоса без облачной маршрутизации.",
   "voice.modelDownload.error": "Загрузка не удалась: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% выполнено",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} МБ из {{total}} МБ",
   "voice.modelDownload.retry": "Повторить попытку",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Скачать модель голоса",
diff --git a/public/locales/sv/bundle.json b/public/locales/sv/bundle.json
index c50e0409..2121d516 100644
--- a/public/locales/sv/bundle.json
+++ b/public/locales/sv/bundle.json
@@ -1805,7 +1805,7 @@
   "settings.ai.localAi.downloadCancelled": "Nedladdningen avbröts",
   "settings.ai.localAi.downloadEta": "ca. {{seconds}}s kvar",
   "settings.ai.localAi.downloadFailed": "Nedladdningen misslyckades. Försök igen.",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB av {{total}} MB",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "Laddar ner AI-modell...",
   "settings.ai.localAi.downloadedLabel": "nedladdade",
@@ -2467,7 +2467,7 @@
   "voice.modelDownload.description": "Laddar ner {{model}}-modellen ({{size}} MB). Detta möjliggör lokal röstbehandling utan molnrouting.",
   "voice.modelDownload.error": "Nedladdning misslyckades: {{error}}",
   "voice.modelDownload.progress": "{{percent}}% klar",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB av {{total}} MB",
   "voice.modelDownload.retry": "Försöka igen",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "Ladda ner röstmodell",
diff --git a/public/locales/zh/bundle.json b/public/locales/zh/bundle.json
index 976f25ba..afada485 100644
--- a/public/locales/zh/bundle.json
+++ b/public/locales/zh/bundle.json
@@ -1805,7 +1805,7 @@
   "settings.ai.localAi.downloadCancelled": "下载已取消",
   "settings.ai.localAi.downloadEta": "约。还剩 {{seconds}} 秒",
   "settings.ai.localAi.downloadFailed": "下载失败。请重试。",
-  "settings.ai.localAi.downloadSize": "{{loaded}} MB of {{total}} MB",
+  "settings.ai.localAi.downloadSize": "{{loaded}} MB／{{total}} MB",
   "settings.ai.localAi.downloadSpeed": "{{speed}} MB/s",
   "settings.ai.localAi.downloadTitle": "正在下载 AI 模型...",
   "settings.ai.localAi.downloadedLabel": "下载的",
@@ -2467,7 +2467,7 @@
   "voice.modelDownload.description": "正在下载 {{model}} 模型 ({{size}} MB)。这使得本地语音处理无需云路由。",
   "voice.modelDownload.error": "下载失败：{{error}}",
   "voice.modelDownload.progress": "{{percent}}% 完成",
-  "voice.modelDownload.progressBytes": "{{loaded}} MB of {{total}} MB",
+  "voice.modelDownload.progressBytes": "{{loaded}} MB／{{total}} MB",
   "voice.modelDownload.retry": "重试",
   "voice.modelDownload.speed": "{{speed}} MB/s",
   "voice.modelDownload.title": "语音模型下载",
diff --git a/services/ai/inferenceProgressEmitter.ts b/services/ai/inferenceProgressEmitter.ts
index 6aa21595..4d698d6c 100644
--- a/services/ai/inferenceProgressEmitter.ts
+++ b/services/ai/inferenceProgressEmitter.ts
@@ -4,6 +4,13 @@
 
 import { WEBLLM_MODEL_APPROX_MB, type WebLlmModelId } from '@domain/ai-core';
 
+// QNBS-v3 (#333/Sourcery): modelId at call sites (services/localAiFacade.ts) is a broader `string` —
+// it can be an adaptive-engine or fallback-chain model id, not just a curated WebLlmModelId — so the
+// param below stays `string`; this guard replaces an unsafe `as WebLlmModelId` cast when indexing.
+function isKnownWebLlmModel(id: string): id is WebLlmModelId {
+  return id in WEBLLM_MODEL_APPROX_MB;
+}
+
 export type WebLlmLoadingState = 'idle' | 'loading' | 'ready' | 'error';
 
 export interface WebLlmLoadProgress {
@@ -64,7 +71,8 @@ class InferenceProgressEmitter {
     const estimatedSecondsRemaining =
       progress > 0.01 && progress < 1 ? Math.round((elapsed / progress) * (1 - progress)) : null;
 
-    const approxMb = modelId ? WEBLLM_MODEL_APPROX_MB[modelId as WebLlmModelId] : undefined;
+    const approxMb =
+      modelId && isKnownWebLlmModel(modelId) ? WEBLLM_MODEL_APPROX_MB[modelId] : undefined;
     const totalBytes = approxMb != null ? approxMb * 1024 * 1024 : null;
     const loadedBytes = totalBytes != null ? Math.round(totalBytes * progress) : null;
     const bytesPerSecond =
diff --git a/services/ai/localModelStorageService.ts b/services/ai/localModelStorageService.ts
index 65a2d445..93236781 100644
--- a/services/ai/localModelStorageService.ts
+++ b/services/ai/localModelStorageService.ts
@@ -3,7 +3,11 @@
 //          the multi-GB downloaded WEIGHTS persist in the Cache API (MLC WebLLM + transformers.js).
 //          This service estimates and clears that on-disk footprint so users can reclaim space.
 
-import { releaseAllOnnxSessions, releaseAllWebLlmEngines } from '@domain/ai-core';
+import {
+  releaseAllOnnxSessions,
+  releaseAllWebLlmEngines,
+  WEBLLM_MODEL_APPROX_MB,
+} from '@domain/ai-core';
 import { logger } from '../logger';
 
 // QNBS-v3: Cache API bucket names used by local-inference runtimes. MLC web-llm shards weights across
@@ -16,17 +20,8 @@ export const LOCAL_MODEL_CACHE_PATTERNS: readonly RegExp[] = [
   /transformers/i,
 ];
 
-// QNBS-v3: Approximate on-disk download size (MB) per curated WebLLM model — drives the pre-download
-//          storage warning. Mirrors the "~X GB" hints in WEBLLM_SUPPORTED_MODELS labels.
-export const WEBLLM_MODEL_APPROX_MB: Readonly<Record<string, number>> = {
-  'Qwen2.5-0.5B-Instruct-q4f16_1-MLC': 400,
-  'Llama-3.2-1B-Instruct-q4f16_1-MLC': 700,
-  'Llama-3.2-3B-Instruct-q4f16_1-MLC': 1800,
-  'Phi-4-mini-instruct-q4f16_1-MLC': 2300,
-  'gemma-3-1b-it-q4f16_1-MLC': 800,
-  'gemma-3-4b-it-q4f32_1-MLC': 4900,
-  'Llama-3.3-70B-Instruct-q3f16_1-MLC': 35_000,
-};
+// QNBS-v3 (#333/Qodo): re-export the canonical @domain/ai-core table instead of maintaining a second, driftable literal — this was a duplicate of the size table now in packages/ai-core/src/index.ts.
+export { WEBLLM_MODEL_APPROX_MB };
 
 export interface LocalModelStorageEstimate {
   // QNBS-v3: null when the StorageManager quota is unknown — callers MUST NOT treat that as "0 free",
diff --git a/services/downloadProgressFormat.ts b/services/downloadProgressFormat.ts
index d45bab88..ca5ee48a 100644
--- a/services/downloadProgressFormat.ts
+++ b/services/downloadProgressFormat.ts
@@ -7,7 +7,10 @@ export function formatMegabytes(bytes: number): string {
   return (bytes / (1024 * 1024)).toFixed(0);
 }
 
-/** Formats a bytes-per-second rate as megabytes/sec with one decimal, e.g. `"3.2"`. */
-export function formatMegabytesPerSecond(bytesPerSecond: number): string {
-  return (bytesPerSecond / (1024 * 1024)).toFixed(1);
+// QNBS-v3 (#333/CodeRabbit): returns a number, not a locale-formatted string — toFixed() always
+// emits a period decimal separator, wrong for comma-decimal locales. Callers format the result with
+// formatNumber() from useTranslation() so the decimal separator matches the active UI locale.
+/** Converts a bytes-per-second rate to megabytes/sec, rounded to one decimal place. */
+export function megabytesPerSecond(bytesPerSecond: number): number {
+  return Math.round((bytesPerSecond / (1024 * 1024)) * 10) / 10;
 }
diff --git a/services/voice/voiceCommandService.ts b/services/voice/voiceCommandService.ts
index 3273a53b..25b087b2 100644
--- a/services/voice/voiceCommandService.ts
+++ b/services/voice/voiceCommandService.ts
@@ -585,9 +585,14 @@ export class VoiceCommandService {
       if (signal?.aborted) return;
 
       // Report initial progress
+      // QNBS-v3 (#333/CodeAnt+Qodo): clear stale byte counts from a previous download attempt —
+      // setVoiceSettings merges partial state, so a prior attempt's bytes would otherwise linger
+      // until the first byte-bearing progress tick of THIS attempt arrives.
       this.d(
         settingsActions.setVoiceSettings({
           wasmModelDownloadProgress: 0.1,
+          wasmModelDownloadLoadedBytes: undefined,
+          wasmModelDownloadTotalBytes: undefined,
         }),
       );
 
@@ -639,11 +644,13 @@ export class VoiceCommandService {
         return;
       }
 
-      // Mark complete
+      // Mark complete — clear byte counts too (no longer meaningful once the download is done).
       this.d(
         settingsActions.setVoiceSettings({
           wasmModelDownloadProgress: 1.0,
           wasmModelsReady: true,
+          wasmModelDownloadLoadedBytes: undefined,
+          wasmModelDownloadTotalBytes: undefined,
         }),
       );
 
@@ -660,6 +667,8 @@ export class VoiceCommandService {
         settingsActions.setVoiceSettings({
           wasmModelDownloadProgress: 0,
           voiceWasmDownloadError: err instanceof Error ? err.message : String(err),
+          wasmModelDownloadLoadedBytes: undefined,
+          wasmModelDownloadTotalBytes: undefined,
         }),
       );
       throw err;
diff --git a/tests/unit/LocalAiDownloadProgress.test.tsx b/tests/unit/LocalAiDownloadProgress.test.tsx
index 955208bc..e7408da5 100644
--- a/tests/unit/LocalAiDownloadProgress.test.tsx
+++ b/tests/unit/LocalAiDownloadProgress.test.tsx
@@ -58,6 +58,8 @@ vi.mock('../../hooks/useTranslation', () => ({
   useTranslation: () => ({
     t: (k: string, args?: Record<string, string>) => (args ? `${k}:${JSON.stringify(args)}` : k),
     language: 'en',
+    formatNumber: (value: number, options?: Intl.NumberFormatOptions) =>
+      value.toFixed(options?.maximumFractionDigits ?? 0),
   }),
 }));
 
@@ -250,10 +252,10 @@ describe('LocalAiDownloadProgress', () => {
     };
     render(<LocalAiDownloadProgress />);
     expect(
-      screen.getByText('settings.ai.localAi.downloadSize:{"loaded":"350","total":"700"}'),
+      screen.getByText('settings.ai.localAi.downloadSize:{"loaded":"~350","total":"~700"}'),
     ).toBeInTheDocument();
     expect(
-      screen.getByText('settings.ai.localAi.downloadSpeed:{"speed":"3.2"}'),
+      screen.getByText('settings.ai.localAi.downloadSpeed:{"speed":"~3.2"}'),
     ).toBeInTheDocument();
   });
 
diff --git a/tests/unit/components/voice/VoiceModelDownloadModal.test.tsx b/tests/unit/components/voice/VoiceModelDownloadModal.test.tsx
index 783c72ef..4536a387 100644
--- a/tests/unit/components/voice/VoiceModelDownloadModal.test.tsx
+++ b/tests/unit/components/voice/VoiceModelDownloadModal.test.tsx
@@ -3,7 +3,7 @@
  * QNBS-v3: P1 tests for P0-5 voice WASM download UI.
  */
 
-import { fireEvent, render, screen, waitFor } from '@testing-library/react';
+import { act, fireEvent, render, screen, waitFor } from '@testing-library/react';
 import type React from 'react';
 import { beforeEach, describe, expect, it, vi } from 'vitest';
 import { VoiceModelDownloadModal } from '../../../../components/voice/VoiceModelDownloadModal';
@@ -32,6 +32,9 @@ vi.mock('../../../../hooks/useTranslation', () => ({
       }
       return template;
     },
+    // QNBS-v3 (#333/Sourcery): mirrors formatNumber's real 1-decimal rendering for the speed display.
+    formatNumber: (value: number, options?: Intl.NumberFormatOptions) =>
+      value.toFixed(options?.maximumFractionDigits ?? 0),
   }),
 }));
 
@@ -187,5 +190,31 @@ describe('VoiceModelDownloadModal', () => {
       await waitFor(() => expect(screen.getByTestId('progress')).toBeInTheDocument());
       expect(screen.queryByText(/MB of/)).not.toBeInTheDocument();
     });
+
+    // QNBS-v3 (#333/Sourcery): exercises the delta-based bytesPerSecond effect end-to-end. Uses fake
+    // timers directly (not waitFor's own polling, which deadlocks once Date.now/setTimeout are faked)
+    // — render/rerender are act()-wrapped by RTL, so state settles synchronously after each call. The
+    // component is React.memo-wrapped, so the second render passes a fresh onClose reference — a
+    // shallow-equal props object would make memo bail out and never re-read the mutated selector value.
+    it('shows a MB/s speed label once two byte samples have elapsed', () => {
+      vi.useFakeTimers();
+      try {
+        mockVoiceState.wasmModelDownloadLoadedBytes = 10 * 1024 * 1024;
+        mockVoiceState.wasmModelDownloadTotalBytes = 42 * 1024 * 1024;
+        const { rerender } = render(<VoiceModelDownloadModal {...defaultProps} />);
+        expect(screen.getByTestId('progress')).toBeInTheDocument();
+        expect(screen.queryByText(/MB\/s/)).not.toBeInTheDocument();
+
+        act(() => {
+          vi.advanceTimersByTime(1000);
+        });
+        mockVoiceState.wasmModelDownloadLoadedBytes = 13.2 * 1024 * 1024;
+        rerender(<VoiceModelDownloadModal {...defaultProps} onClose={vi.fn()} />);
+
+        expect(screen.getByText('3.2 MB/s')).toBeInTheDocument();
+      } finally {
+        vi.useRealTimers();
+      }
+    });
   });
 });
diff --git a/tests/unit/services/downloadProgressFormat.test.ts b/tests/unit/services/downloadProgressFormat.test.ts
index bdc4b083..46a25146 100644
--- a/tests/unit/services/downloadProgressFormat.test.ts
+++ b/tests/unit/services/downloadProgressFormat.test.ts
@@ -1,8 +1,5 @@
 import { describe, expect, it } from 'vitest';
-import {
-  formatMegabytes,
-  formatMegabytesPerSecond,
-} from '../../../services/downloadProgressFormat';
+import { formatMegabytes, megabytesPerSecond } from '../../../services/downloadProgressFormat';
 
 describe('formatMegabytes', () => {
   it('formats bytes as whole megabytes', () => {
@@ -18,12 +15,17 @@ describe('formatMegabytes', () => {
   });
 });
 
-describe('formatMegabytesPerSecond', () => {
-  it('formats a byte rate with one decimal', () => {
-    expect(formatMegabytesPerSecond(3.2 * 1024 * 1024)).toBe('3.2');
+describe('megabytesPerSecond', () => {
+  it('converts a byte rate to megabytes, rounded to one decimal', () => {
+    expect(megabytesPerSecond(3.2 * 1024 * 1024)).toBe(3.2);
   });
 
   it('handles zero bytes per second', () => {
-    expect(formatMegabytesPerSecond(0)).toBe('0.0');
+    expect(megabytesPerSecond(0)).toBe(0);
+  });
+
+  // QNBS-v3 (#333/CodeRabbit): the caller — not this function — is responsible for locale-aware decimal formatting.
+  it('returns a plain number, not a locale-formatted string', () => {
+    expect(typeof megabytesPerSecond(1024 * 1024)).toBe('number');
   });
 });
diff --git a/tests/unit/services/voice/voiceCommandService.test.ts b/tests/unit/services/voice/voiceCommandService.test.ts
index bd982b7f..9a18445d 100644
--- a/tests/unit/services/voice/voiceCommandService.test.ts
+++ b/tests/unit/services/voice/voiceCommandService.test.ts
@@ -327,5 +327,77 @@ describe('VoiceCommandService', () => {
         );
       expect(progressCalls.some((a) => a.payload.wasmModelDownloadProgress === 0.3)).toBe(true);
     });
+
+    // QNBS-v3 (#333/CodeAnt+Qodo+CodeRabbit): byte fields must be explicitly cleared at every
+    // lifecycle boundary — setVoiceSettings merges partial state, so an omitted key would leave a
+    // prior attempt's bytes visible until the next byte-bearing progress tick arrives.
+    it('clears stale byte counts from a previous attempt at the start of a new download', async () => {
+      mockPipeline.mockImplementation(async (_task: string, _model: string, opts: unknown) => {
+        const { onProgress } = opts as { onProgress: (p: unknown) => void };
+        onProgress({ progress: 10, loaded: 1, total: 10 });
+        return { dispose: vi.fn() };
+      });
+
+      await service.downloadVoiceModels('stt');
+
+      const initialCall = dispatch.mock.calls
+        .map((c) => c[0])
+        .find(
+          (a) =>
+            (a as { payload?: { wasmModelDownloadProgress?: number } })?.payload
+              ?.wasmModelDownloadProgress === 0.1,
+        ) as { payload: Record<string, unknown> } | undefined;
+      expect(initialCall).toBeDefined();
+      expect(initialCall?.payload).toHaveProperty('wasmModelDownloadLoadedBytes');
+      expect(initialCall?.payload['wasmModelDownloadLoadedBytes']).toBeUndefined();
+      expect(initialCall?.payload).toHaveProperty('wasmModelDownloadTotalBytes');
+      expect(initialCall?.payload['wasmModelDownloadTotalBytes']).toBeUndefined();
+    });
+
+    it('clears byte counts on successful completion', async () => {
+      mockPipeline.mockImplementation(async (_task: string, _model: string, opts: unknown) => {
+        const { onProgress } = opts as { onProgress: (p: unknown) => void };
+        onProgress({ progress: 50, loaded: 21_000_000, total: 42_000_000 });
+        return { dispose: vi.fn() };
+      });
+
+      await service.downloadVoiceModels('stt');
+
+      const completeCall = dispatch.mock.calls
+        .map((c) => c[0])
+        .find(
+          (a) =>
+            (a as { payload?: { wasmModelDownloadProgress?: number } })?.payload
+              ?.wasmModelDownloadProgress === 1.0,
+        ) as { payload: Record<string, unknown> } | undefined;
+      expect(completeCall).toBeDefined();
+      expect(completeCall?.payload).toHaveProperty('wasmModelDownloadLoadedBytes');
+      expect(completeCall?.payload['wasmModelDownloadLoadedBytes']).toBeUndefined();
+      expect(completeCall?.payload).toHaveProperty('wasmModelDownloadTotalBytes');
+      expect(completeCall?.payload['wasmModelDownloadTotalBytes']).toBeUndefined();
+    });
+
+    it('clears byte counts when the download fails', async () => {
+      mockPipeline.mockImplementation(async (_task: string, _model: string, opts: unknown) => {
+        const { onProgress } = opts as { onProgress: (p: unknown) => void };
+        onProgress({ progress: 50, loaded: 21_000_000, total: 42_000_000 });
+        throw new Error('network error');
+      });
+
+      await expect(service.downloadVoiceModels('stt')).rejects.toThrow('network error');
+
+      const errorCall = dispatch.mock.calls
+        .map((c) => c[0])
+        .find(
+          (a) =>
+            (a as { payload?: { voiceWasmDownloadError?: string } })?.payload
+              ?.voiceWasmDownloadError === 'network error',
+        ) as { payload: Record<string, unknown> } | undefined;
+      expect(errorCall).toBeDefined();
+      expect(errorCall?.payload).toHaveProperty('wasmModelDownloadLoadedBytes');
+      expect(errorCall?.payload['wasmModelDownloadLoadedBytes']).toBeUndefined();
+      expect(errorCall?.payload).toHaveProperty('wasmModelDownloadTotalBytes');
+      expect(errorCall?.payload['wasmModelDownloadTotalBytes']).toBeUndefined();
+    });
   });
 });
diff --git a/types.ts b/types.ts
index b9bc5f7d..94f21a2b 100644
--- a/types.ts
+++ b/types.ts
@@ -611,10 +611,11 @@ export interface VoiceSettings {
   wasmModelDownloadProgress?: number;
   /** QNBS-v3 (#333 item 1): real bytes loaded so far, from transformers.js's own progress payload
    * (not an approximation, unlike the WebLLM text-model path — this download is a raw fetch the
-   * library already reports byte counts for). Undefined until the first progress tick arrives. */
-  wasmModelDownloadLoadedBytes?: number;
+   * library already reports byte counts for). Undefined until the first progress tick arrives, and
+   * explicitly cleared (not just omitted) at every download lifecycle boundary — see downloadVoiceModels. */
+  wasmModelDownloadLoadedBytes?: number | undefined;
   /** QNBS-v3 (#333 item 1): real total byte size of the file currently downloading. */
-  wasmModelDownloadTotalBytes?: number;
+  wasmModelDownloadTotalBytes?: number | undefined;
   /** WASM voice models ready for use */
   wasmModelsReady?: boolean;
   /** WASM voice model download error message */