ggml-org · allozaur · Mar 31, 2026 · Mar 27, 2026 · Mar 27, 2026 · Mar 27, 2026
diff --git a/.editorconfig b/.editorconfig
@@ -21,14 +21,6 @@ indent_style = tab
 [prompts/*.txt]
 insert_final_newline = unset
 
-[tools/server/public/*]
-indent_size = 2
-
-[tools/server/public/deps_*]
-trim_trailing_whitespace = unset
-indent_style = unset
-indent_size = unset
-
 [tools/server/deps_*]
 trim_trailing_whitespace = unset
 indent_style = unset
@@ -61,6 +53,14 @@ charset = unset
 trim_trailing_whitespace = unset
 insert_final_newline = unset
 
+[tools/server/public/**]
+indent_style = unset
+indent_size = unset
+end_of_line = unset
+charset = unset
+trim_trailing_whitespace = unset
+insert_final_newline = unset
+
 [benches/**]
 indent_style = unset
 indent_size = unset

diff --git a/.gitattributes b/.gitattributes
@@ -0,0 +1,4 @@
+# Treat the generated single-file WebUI build as binary for diff purposes.
+# Git's pack-file delta compression still works (byte-level), but this prevents
+# git diff from printing the entire minified file on every change.
+tools/server/public/index.html -diff
diff --git a/.gitignore b/.gitignore
@@ -95,6 +95,8 @@
 # Server Web UI temporary files
 /tools/server/webui/node_modules
 /tools/server/webui/dist
+# we no longer use gz for index.html
+/tools/server/public/index.html.gz
 
 # Python
 

@@ -42,7 +42,9 @@ option(LLAMA_BUILD_WEBUI "Build the embedded Web UI" ON)
 
 if (LLAMA_BUILD_WEBUI)
     set(PUBLIC_ASSETS
-        index.html.gz
+        index.html
+        bundle.js
+        bundle.css
         loading.html
     )
 

@@ -259,6 +259,6 @@ npm run test
 npm run build
 ```
 
-After `public/index.html.gz` has been generated, rebuild `llama-server` as described in the [build](#build) section to include the updated UI.
+After `public/index.html` has been generated, rebuild `llama-server` as described in the [build](#build) section to include the updated UI.
 
 **Note:** The Vite dev server automatically proxies API requests to `http://localhost:8080`. Make sure `llama-server` is running on that port during development.
diff --git a/tools/server/public/bundle.css b/tools/server/public/bundle.css
diff --git a/tools/server/public/bundle.js b/tools/server/public/bundle.js
diff --git a/tools/server/public/index.html b/tools/server/public/index.html
@@ -0,0 +1,34 @@
+<!--
+  This is a single file build of the frontend.
+  It is automatically generated by the build process.
+  Do not edit this file directly.
+  To make changes, refer to the "Web UI" section in the README.
+-->
+<!doctype html>
+<html lang="en">
+	<head>
+		<meta charset="utf-8" />
+		<link rel="icon" href="data:image/svg+xml;base64,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" />
+		<meta name="viewport" content="width=device-width, initial-scale=1" />
+
+		<link href="./bundle.css" rel="stylesheet">
+		<link rel="modulepreload" href="./bundle.js">
+	</head>
+	<body data-sveltekit-preload-data="hover">
+		<div style="display: contents">
+			<script>
+				{
+					__sveltekit_286ovp = {
+						base: new URL('.', location).pathname.slice(0, -1)
+					};
+
+					const element = document.currentScript.parentElement;
+
+					import("./bundle.js").then((app) => {
+						app.start(element)
+					});
+				}
+			</script>
+		</div>
+	</body>
+</html>
diff --git a/tools/server/public/index.html.gz b/tools/server/public/index.html.gz
@@ -10,7 +10,9 @@
 
 #ifdef LLAMA_BUILD_WEBUI
 // auto generated files (see README.md for details)
-#include "index.html.gz.hpp"
+#include "index.html.hpp"
+#include "bundle.js.hpp"
+#include "bundle.css.hpp"
 #include "loading.html.hpp"
 #endif
 
@@ -262,16 +264,19 @@ bool server_http_context::init(const common_params & params) {
         } else {
 #ifdef LLAMA_BUILD_WEBUI
             // using embedded static index.html
-            srv->Get(params.api_prefix + "/", [](const httplib::Request & req, httplib::Response & res) {
-                if (req.get_header_value("Accept-Encoding").find("gzip") == std::string::npos) {
-                    res.set_content("Error: gzip is not supported by this browser", "text/plain");
-                } else {
-                    res.set_header("Content-Encoding", "gzip");
-                    // COEP and COOP headers, required by pyodide (python interpreter)
-                    res.set_header("Cross-Origin-Embedder-Policy", "require-corp");
-                    res.set_header("Cross-Origin-Opener-Policy", "same-origin");
-                    res.set_content(reinterpret_cast<const char*>(index_html_gz), index_html_gz_len, "text/html; charset=utf-8");
-                }
+            srv->Get(params.api_prefix + "/", [](const httplib::Request & /*req*/, httplib::Response & res) {
+                // COEP and COOP headers, required by pyodide (python interpreter)
+                res.set_header("Cross-Origin-Embedder-Policy", "require-corp");
+                res.set_header("Cross-Origin-Opener-Policy", "same-origin");
+                res.set_content(reinterpret_cast<const char*>(index_html), index_html_len, "text/html; charset=utf-8");
+                return false;
+            });
+            srv->Get(params.api_prefix + "/bundle.js", [](const httplib::Request & /*req*/, httplib::Response & res) {
+                res.set_content(reinterpret_cast<const char*>(bundle_js), bundle_js_len, "application/javascript; charset=utf-8");
+                return false;
+            });
+            srv->Get(params.api_prefix + "/bundle.css", [](const httplib::Request & /*req*/, httplib::Response & res) {
+                res.set_content(reinterpret_cast<const char*>(bundle_css), bundle_css_len, "text/css; charset=utf-8");
                 return false;
             });
 #endif

diff --git a/tools/server/webui/README.md b/tools/server/webui/README.md
@@ -188,14 +188,14 @@ The build process:
 1. **Vite Build** - Bundles all TypeScript, Svelte, and CSS
 2. **Static Adapter** - Outputs to `../public` (llama-server's static file directory)
 3. **Post-Build Script** - Cleans up intermediate files
-4. **Custom Plugin** - Creates `index.html.gz` with:
+4. **Custom Plugin** - Creates `index.html` with:
    - Inlined favicon as base64
    - GZIP compression (level 9)
    - Deterministic output (zeroed timestamps)
 
 ```text
 tools/server/webui/        →  build  →  tools/server/public/
-├── src/                                 ├── index.html.gz  (served by llama-server)
+├── src/                                 ├── index.html  (served by llama-server)
 ├── static/                              └── (favicon inlined)
 └── ...
 ```
@@ -219,7 +219,7 @@ output: {
 
 The WebUI is embedded directly into the llama-server binary:
 
-1. `npm run build` outputs `index.html.gz` to `tools/server/public/`
+1. `npm run build` outputs `index.html` to `tools/server/public/`
 2. llama-server compiles this into the binary at build time
 3. When accessing `/`, llama-server serves the gzipped HTML
 4. All assets are inlined (CSS, JS, fonts, favicon)

diff --git a/tools/server/webui/package.json b/tools/server/webui/package.json
@@ -50,7 +50,6 @@
 		"eslint-config-prettier": "^10.0.1",
 		"eslint-plugin-storybook": "^10.2.4",
 		"eslint-plugin-svelte": "^3.0.0",
-		"fflate": "^0.8.2",
 		"globals": "^16.0.0",
 		"http-server": "^14.1.1",
 		"mdast": "^3.0.0",

diff --git a/tools/server/webui/scripts/install-git-hooks.sh b/tools/server/webui/scripts/install-git-hooks.sh
@@ -83,7 +83,7 @@ if [ -n "$WEBUI_CHANGES" ]; then
     fi
 
     # Check if build output exists and is newer than source files
-    BUILD_FILE="../public/index.html.gz"
+    BUILD_FILE="../public/index.html"
     NEEDS_BUILD=false
 
     if [ ! -f "$BUILD_FILE" ]; then
@@ -127,9 +127,9 @@ if [ -n "$WEBUI_CHANGES" ]; then
         cd ../../..
 
         # Check if build output was created/updated
-        if [ -f "tools/server/public/index.html.gz" ]; then
+        if [ -f "tools/server/public/index.html" ]; then
             # Add the build output and commit it
-            git add tools/server/public/index.html.gz
+            git add tools/server/public/index.html
             if ! git diff --cached --quiet; then
                 echo "Committing updated build output..."
                 git commit -m "chore: update webui build output"

diff --git a/tools/server/webui/scripts/post-build.sh b/tools/server/webui/scripts/post-build.sh
@@ -1,3 +1,3 @@
 rm -rf ../public/_app;
 rm ../public/favicon.svg;
-rm ../public/index.html;
+rm -f ../public/index.html.gz; # deprecated, but may still be generated by older versions of the build process
diff --git a/tools/server/webui/svelte.config.js b/tools/server/webui/svelte.config.js
@@ -21,7 +21,7 @@ const config = {
 			strict: true
 		}),
 		output: {
-			bundleStrategy: 'inline'
+			bundleStrategy: 'single'
 		},
 		alias: {
 			$styles: 'src/styles'

diff --git a/tools/server/webui/tests/e2e/demo.test.ts b/tools/server/webui/tests/e2e/demo.test.ts
@@ -2,5 +2,5 @@ import { expect, test } from '@playwright/test';
 
 test('home page has expected h1', async ({ page }) => {
 	await page.goto('/');
-	await expect(page.locator('h1')).toBeVisible();
+	await expect(page.locator('h1').first()).toBeVisible();
 });
diff --git a/tools/server/webui/vite.config.ts b/tools/server/webui/vite.config.ts
@@ -1,7 +1,6 @@
 import tailwindcss from '@tailwindcss/vite';
 import { sveltekit } from '@sveltejs/kit/vite';
-import * as fflate from 'fflate';
-import { readFileSync, writeFileSync, existsSync } from 'fs';
+import { readFileSync, writeFileSync, existsSync, readdirSync, copyFileSync } from 'fs';
 import { dirname, resolve } from 'path';
 import { fileURLToPath } from 'url';
 
@@ -20,15 +19,13 @@ const GUIDE_FOR_FRONTEND = `
 -->
 `.trim();
 
-const MAX_BUNDLE_SIZE = 2 * 1024 * 1024;
-
 /**
  * the maximum size of an embedded asset in bytes,
  * e.g. maximum size of embedded font (see node_modules/katex/dist/fonts/*.woff2)
  */
 const MAX_ASSET_SIZE = 32000;
 
-/** public/index.html.gz minified flag */
+/** public/index.html minified flag */
 const ENABLE_JS_MINIFICATION = true;
 
 function llamaCppBuildPlugin() {
@@ -40,7 +37,6 @@ function llamaCppBuildPlugin() {
 			setTimeout(() => {
 				try {
 					const indexPath = resolve('../public/index.html');
-					const gzipPath = resolve('../public/index.html.gz');
 
 					if (!existsSync(indexPath)) {
 						return;
@@ -61,26 +57,35 @@ function llamaCppBuildPlugin() {
 
 					content = content.replace(/\r/g, '');
 					content = GUIDE_FOR_FRONTEND + '\n' + content;
-
-					const compressed = fflate.gzipSync(Buffer.from(content, 'utf-8'), { level: 9 });
-
-					compressed[0x4] = 0;
-					compressed[0x5] = 0;
-					compressed[0x6] = 0;
-					compressed[0x7] = 0;
-					compressed[0x9] = 0;
-
-					if (compressed.byteLength > MAX_BUNDLE_SIZE) {
-						throw new Error(
-							`Bundle size is too large (${Math.ceil(compressed.byteLength / 1024)} KB).\n` +
-								`Please reduce the size of the frontend or increase MAX_BUNDLE_SIZE in vite.config.ts.\n`
-						);
+					content = content.replace(/\/_app\/immutable\/bundle\.[^"]+\.js/g, './bundle.js');
+					content = content.replace(
+						/\/_app\/immutable\/assets\/bundle\.[^"]+\.css/g,
+						'./bundle.css'
+					);
+
+					writeFileSync(indexPath, content, 'utf-8');
+					console.log('✓ Updated index.html');
+
+					// Copy bundle.*.js -> ../public/bundle.js
+					const immutableDir = resolve('../public/_app/immutable');
+					const bundleDir = resolve('../public/_app/immutable/assets');
+					if (existsSync(immutableDir)) {
+						const jsFiles = readdirSync(immutableDir).filter((f) => f.match(/^bundle\..+\.js$/));
+						if (jsFiles.length > 0) {
+							copyFileSync(resolve(immutableDir, jsFiles[0]), resolve('../public/bundle.js'));
+							console.log(`✓ Copied ${jsFiles[0]} -> bundle.js`);
+						}
+					}
+					// Copy bundle.*.css -> ../public/bundle.css
+					if (existsSync(bundleDir)) {
+						const cssFiles = readdirSync(bundleDir).filter((f) => f.match(/^bundle\..+\.css$/));
+						if (cssFiles.length > 0) {
+							copyFileSync(resolve(bundleDir, cssFiles[0]), resolve('../public/bundle.css'));
+							console.log(`✓ Copied ${cssFiles[0]} -> bundle.css`);
+						}
 					}
-
-					writeFileSync(gzipPath, compressed);
-					console.log('✓ Created index.html.gz');
 				} catch (error) {
-					console.error('Failed to create gzip file:', error);
+					console.error('Failed to update index.html:', error);
 				}
 			}, 100);
 		}