fix(tts): resolve Gemini TTS models from catalog

Resolve Gemini TTS models from shared TTS catalog and provider registry
with a safe fallback, fixing requests resolving to models/undefined when
ttsConfig.models is empty. Add gemini-3.1-flash-tts-preview to catalogs.

Co-authored-by: Cursor <cursoragent@cursor.com>

open-sse/config/ttsModels.js

@@ -96,10 +96,12 @@ export const TTS_MODELS_CONFIG = {
   },
   gemini: {
     models: [
+      { id: "gemini-3.1-flash-tts-preview", name: "Gemini 3.1 Flash TTS", type: "tts" },
       { id: "gemini-2.5-flash-preview-tts", name: "Gemini 2.5 Flash TTS", type: "tts" },
       { id: "gemini-2.5-pro-preview-tts",   name: "Gemini 2.5 Pro TTS",   type: "tts" },
     ],
     voices: {
+      "gemini-3.1-flash-tts-preview": GEMINI_VOICES,
       "gemini-2.5-flash-preview-tts": GEMINI_VOICES,
       "gemini-2.5-pro-preview-tts":   GEMINI_VOICES,
     },

open-sse/handlers/ttsProviders/gemini.js

@@ -1,11 +1,19 @@
 // Gemini TTS — generateContent with AUDIO modality returns PCM L16, wrap as WAV
 import { Buffer } from "node:buffer";
-import { PROVIDER_MEDIA } from "../../providers/index.js";
+import { PROVIDER_MEDIA, PROVIDER_MODELS } from "../../providers/index.js";
 
 const TTS_CFG = PROVIDER_MEDIA["gemini"]?.ttsConfig || {};
 const TTS_BASE = TTS_CFG.baseUrl;
-const KNOWN_MODELS = (TTS_CFG.models || []).map((m) => m.id);
-const DEFAULT_MODEL = KNOWN_MODELS[0];
+const FALLBACK_MODEL = "gemini-3.1-flash-tts-preview";
+const KNOWN_MODELS = [
+  ...(TTS_CFG.models || []),
+  ...(PROVIDER_MODELS["gemini-tts-models"] || []),
+  ...(PROVIDER_MODELS.gemini || []).filter((m) => (m.kind || m.type) === "tts"),
+]
+  .map((m) => m?.id)
+  .filter(Boolean)
+  .filter((id, index, list) => list.indexOf(id) === index);
+const DEFAULT_MODEL = KNOWN_MODELS[0] || FALLBACK_MODEL;
 const DEFAULT_VOICE = "Kore";
 
 // Parse "model/voice" — if input doesn't match a known TTS model, treat it as voice with default model

open-sse/providers/registry/gemini.js

@@ -52,6 +52,7 @@ export default {
     { id: "gemini-2.5-flash", name: "Gemini 2.5 Flash", params: ["language","prompt"], kind: "stt" },
     { id: "gemini-2.5-flash-lite", name: "Gemini 2.5 Flash Lite (Cheapest)", params: ["language","prompt"], kind: "stt" },
     { id: "gemini-2.0-flash", name: "Gemini 2.0 Flash", params: ["language","prompt"], kind: "stt" },
+    { id: "gemini-3.1-flash-tts-preview", name: "Gemini 3.1 Flash TTS", kind: "tts" },
     { id: "gemini-2.5-flash-preview-tts", name: "Gemini 2.5 Flash TTS", kind: "tts" },
     { id: "gemini-2.5-pro-preview-tts", name: "Gemini 2.5 Pro TTS", kind: "tts" },
     { id: "embedding-001", name: "Embedding 001", dimensions: 768, kind: "embedding" },

tests/unit/gemini-tts.test.js

@@ -0,0 +1,95 @@
+import { describe, it, expect, vi, beforeEach, afterEach } from "vitest";
+import { handleTtsCore } from "../../open-sse/handlers/ttsCore.js";
+import { buildTtsProviderModels } from "../../open-sse/config/ttsModels.js";
+
+const originalFetch = global.fetch;
+
+function mockGeminiAudioResponse() {
+  global.fetch.mockResolvedValueOnce(
+    new Response(
+      JSON.stringify({
+        candidates: [
+          {
+            content: {
+              parts: [
+                {
+                  inlineData: {
+                    mimeType: "audio/pcm",
+                    data: Buffer.from([0, 1, 2, 3]).toString("base64"),
+                  },
+                },
+              ],
+            },
+          },
+        ],
+      }),
+      { status: 200, headers: { "Content-Type": "application/json" } }
+    )
+  );
+}
+
+describe("Gemini TTS", () => {
+  beforeEach(() => {
+    global.fetch = vi.fn();
+  });
+
+  afterEach(() => {
+    global.fetch = originalFetch;
+  });
+
+  it("uses the default Gemini TTS model when only a voice is provided", async () => {
+    mockGeminiAudioResponse();
+
+    const result = await handleTtsCore({
+      provider: "gemini",
+      model: "Zephyr",
+      input: "Hello from Gemini",
+      credentials: { apiKey: "test-key" },
+      responseFormat: "json",
+    });
+
+    expect(result.success).toBe(true);
+    expect(global.fetch.mock.calls[0][0]).toBe(
+      "https://generativelanguage.googleapis.com/v1beta/models/gemini-3.1-flash-tts-preview:generateContent?key=test-key"
+    );
+
+    const sent = JSON.parse(global.fetch.mock.calls[0][1].body);
+    expect(sent.generationConfig.speechConfig.voiceConfig.prebuiltVoiceConfig.voiceName).toBe("Zephyr");
+    const body = await result.response.json();
+    expect(body.format).toBe("wav");
+    expect(body.audio).toEqual(expect.any(String));
+  });
+
+  it("preserves an explicit Gemini TTS model and voice pair", async () => {
+    mockGeminiAudioResponse();
+
+    const result = await handleTtsCore({
+      provider: "gemini",
+      model: "gemini-2.5-flash-preview-tts/Puck",
+      input: "Hello from Gemini",
+      credentials: { apiKey: "test-key" },
+      responseFormat: "json",
+    });
+
+    expect(result.success).toBe(true);
+    expect(global.fetch.mock.calls[0][0]).toBe(
+      "https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-preview-tts:generateContent?key=test-key"
+    );
+
+    const sent = JSON.parse(global.fetch.mock.calls[0][1].body);
+    expect(sent.generationConfig.speechConfig.voiceConfig.prebuiltVoiceConfig.voiceName).toBe("Puck");
+  });
+
+  it("exposes current Gemini TTS models in the TTS catalog", () => {
+    const entries = buildTtsProviderModels();
+
+    expect(entries["gemini-tts-models"].map((model) => model.id)).toEqual([
+      "gemini-3.1-flash-tts-preview",
+      "gemini-2.5-flash-preview-tts",
+      "gemini-2.5-pro-preview-tts",
+    ]);
+    expect(entries["gemini-tts-voices"]).toContainEqual(
+      expect.objectContaining({ id: "Zephyr", type: "tts" })
+    );
+  });
+});