@agentprojectcontext/apx 1.65.2 → 1.65.3

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@agentprojectcontext/apx",
-  "version": "1.65.2",
+  "version": "1.65.3",
   "description": "APX — unified CLI + daemon for the Agent Project Context (APC) standard.",
   "publishConfig": {
     "access": "public"

package/src/core/voice/engines/gemini.js

@@ -1,13 +1,12 @@
 // Gemini TTS adapter.
-// At the time of writing the Gemini TTS surface is gated and not stable across
-// SDK versions: some models (gemini-2.5-flash-preview-tts) expose synthesize
-// via the v1beta REST surface, others require Vertex. To keep APX engine-
-// agnostic, this adapter performs a best-effort call against the documented
-// REST shape, but flags itself as not-implemented when the response does not
-// include inline audio data.
+// The Gemini TTS surface can vary across SDK versions: some models
+// (gemini-2.5-flash-tts) expose synthesize via the v1beta REST surface, others
+// require Vertex. To keep APX engine-agnostic, this adapter performs a
+// best-effort call against the documented REST shape, but flags itself as
+// not-implemented when the response does not include inline audio data.
 //
 // Config (~/.apx/config.json → voice.tts.gemini):
-//   { "api_key": "...", "model": "gemini-2.5-flash-preview-tts", "voice": "Kore",
+//   { "api_key": "...", "model": "gemini-2.5-flash-tts", "voice": "Kore",
 //     "style": "habla en tono alegre y enérgico" }
 //
 // `style` is an optional natural-language instruction describing HOW the voice
@@ -22,7 +21,7 @@ import fs from "node:fs";
 import path from "node:path";
 import { randomUUID } from "node:crypto";
 
-const DEFAULT_MODEL = "gemini-2.5-flash-preview-tts";
+const DEFAULT_MODEL = "gemini-2.5-flash-tts";
 
 function getKey(config, parentEnginesCfg) {
   return (

package/src/interfaces/cli/commands/setup.js

@@ -88,7 +88,7 @@ const PROVIDERS = [
     needsKey: true,
     keyLabel: "Gemini API key",
     keyHint: "AIza...",
-    models: ["gemini-2.0-flash", "gemini-1.5-pro"],
+    models: ["gemini-3.5-flash", "gemini-3.1-pro-preview", "gemini-2.5-flash"],
   },
 ];
 
@@ -254,7 +254,7 @@ export async function cmdSetup() {
     console.log(`    ${cy("2")}. piper      ${di("(local, offline; needs piper CLI + voice model)")}`);
     console.log(`    ${cy("3")}. elevenlabs ${di("(cloud; eleven_multilingual_v2)")}`);
     console.log(`    ${cy("4")}. openai     ${di("(cloud; tts-1, reuses your openai key)")}`);
-    console.log(`    ${cy("5")}. gemini     ${di("(experimental, best-effort)")}`);
+    console.log(`    ${cy("5")}. gemini     ${di("(cloud; gemini-2.5-flash-tts, 30 voices)")}`);
     console.log(`    ${cy("6")}. mock       ${di("(silent WAV; useful for tests)")}`);
     console.log();
     const choice = (await ask(`  Choose [1-6, default 1]: `)).trim() || "1";