feat(providers): optional AI Horde API key and live image catalog

changelog.d/features/10542-aihorde-optional-key-image-catalog.md

@@ -0,0 +1,2 @@
+- **feat(providers):** AI Horde accepts an optional registered API key and advertises only live image models that currently have workers ([#10542](https://github.com/diegosouzapw/OmniRoute/pull/10542))
+- **fix(providers):** AI Horde Check validates keys via `/v2/find_user` instead of the unauthenticated OpenAI models list ([#10542](https://github.com/diegosouzapw/OmniRoute/pull/10542))

open-sse/config/imageRegistry.ts

@@ -16,6 +16,7 @@ import {
   ADOBE_FIREFLY_IMAGE_ROUTING_ALIASES,
   toRegistryImageModels,
 } from "../services/adobeFireflyModels.ts";
+import { AI_HORDE_IMAGE_PROVIDER } from "./providers/registry/aihorde/imageModels.ts";
 
 interface ImageModelEntry {
   id: string;
@@ -841,6 +842,7 @@ export const IMAGE_PROVIDERS: Record<string, ImageProviderConfig> = {
     // still pass supported 4K dimensions through the permissive request schema.
     supportedSizes: ["1024x1024", "2048x2048"],
   },
+  aihorde: AI_HORDE_IMAGE_PROVIDER,
 };
 
 /**

open-sse/config/providers/registry/aihorde/imageModels.ts

@@ -0,0 +1,26 @@
+/**
+ * AI Horde image-generation provider entry.
+ *
+ * Chat still goes through oai.aihorde.net. Image jobs use the native Horde
+ * async API (`/v2/generate/async`). `models` is a live getter so
+ * imageRegistry stays under the file-size cap and zero-worker names are
+ * never advertised.
+ */
+import { getCachedAiHordeImageCatalogEntries } from "../../../../services/aihordeImageCatalog.ts";
+
+export const AI_HORDE_IMAGE_PROVIDER = {
+  id: "aihorde",
+  alias: "horde",
+  baseUrl: "https://aihorde.net/api",
+  authType: "apikey",
+  authHeader: "apikey",
+  format: "aihorde",
+  get models() {
+    return getCachedAiHordeImageCatalogEntries().map((entry) => ({
+      id: entry.id.startsWith("aihorde/") ? entry.id.slice("aihorde/".length) : entry.id,
+      name: entry.name,
+      inputModalities: entry.inputModalities,
+    }));
+  },
+  supportedSizes: ["512x512", "768x768", "1024x1024", "1024x768", "768x1024"],
+};

open-sse/config/providers/registry/aihorde/index.ts

@@ -17,9 +17,14 @@ import { buildOpenAiCompatibleRegistryEntry } from "../../shared.ts";
  *    the free catalog registers it as `recurring-uncapped` (never summed into
  *    the token headline) rather than inventing an RPM/RPD figure.
  *
- * Model list changes as workers come and go, so the live catalog is fetched via
- * passthrough; the entries below are the ones that have carried steady worker
- * threads and only serve as a fallback when discovery fails.
+ * Chat model list changes as workers come and go, so the live chat catalog is
+ * fetched via passthrough; the entries below are the ones that have carried
+ * steady worker threads and only serve as a fallback when discovery fails.
+ *
+ * Image models are a separate native Horde API (`/v2/generate/async`). They
+ * are discovered by polling `/v2/status/models?type=image` and only advertised
+ * while `count > 0`. An optional registered API key is stored as a normal
+ * connection and sent as the Horde `apikey` header for both chat and images.
  */
 export const aihordeProvider: RegistryEntry = buildOpenAiCompatibleRegistryEntry({
   id: "aihorde",

open-sse/handlers/imageGeneration.ts

@@ -1,20 +1,5 @@
 import { randomUUID } from "crypto";
-/**
- * Image Generation Handler
- *
- * Handles POST /v1/images/generations requests.
- * Proxies to upstream image generation providers using OpenAI-compatible format.
- *
- * Request format (OpenAI-compatible):
- * {
- *   "model": "openai/gpt-image-2",
- *   "prompt": "a beautiful sunset over mountains",
- *   "n": 1,
- *   "size": "1024x1024",
- *   "quality": "standard",       // optional: "standard" | "hd"
- *   "response_format": "url"     // optional: "url" | "b64_json"
- * }
- */
+/** Image generation handler for POST /v1/images/generations (OpenAI-compatible). */
 
 import { getImageProvider, parseImageModel } from "../config/imageRegistry.ts";
 import { HTTP_STATUS } from "../config/constants.ts";
@@ -51,10 +36,6 @@ import {
 } from "@/shared/utils/fetchTimeout";
 import { sanitizeErrorMessage, sanitizeUpstreamDetails } from "../utils/error.ts";
 
-// --- Per-provider handlers (extracted to co-located files in PR-#4582-batch) ---
-// Imported locally so internal callers (handleImageGeneration / handleImageEdit)
-// resolve to a real binding. extractMarkdownImageUrls + CHATGPT_WEB_IMAGE_ID_RE
-// are still used by handleImageEdit below, so they are imported (not re-defined).
 import { handleSDWebUIImageGeneration } from "./imageGeneration/providers/sdWebUI.ts";
 import { handleHyperbolicImageGeneration } from "./imageGeneration/providers/hyperbolic.ts";
 import { handleHuggingFaceImageGeneration } from "./imageGeneration/providers/huggingface.ts";
@@ -76,6 +57,7 @@ import { handleDesignerWebImageGeneration } from "./imageGeneration/providers/de
 import { handleMinimaxImageGeneration } from "./imageGeneration/providers/minimax.ts";
 import { handleAdobeFireflyImageGeneration } from "./imageGeneration/providers/adobeFirefly.ts";
 import { handleAlibabaImageGeneration } from "./imageGeneration/providers/alibabaImage.ts";
+import { handleAiHordeImageGeneration } from "./imageGeneration/providers/aihorde.ts";
 import {
   applyPollinationsAnonymousFallback,
   reportPollinationsAnonOutcome,
@@ -373,6 +355,18 @@ export async function handleImageGeneration({
     });
   }
 
+  if (providerConfig.format === "aihorde") {
+    return handleAiHordeImageGeneration({
+      model,
+      provider,
+      providerConfig,
+      body,
+      credentials,
+      log,
+      signal,
+    });
+  }
+
   if (providerConfig.format === "gemini-image") {
     return handleGeminiImageGeneration({ model, providerConfig, body, credentials, log });
   }