feat(voice): real TTS fallback chain and whisper-compatible STT

- TTS providers: cloud (OpenAI-compatible /audio/speech, e.g. Bailian
  compatible mode with cosyvoice), local self-hosted HTTP server,
  edge-tts CLI, espeak CLI, silent — ordered by TTS_PROVIDER list,
  skipping providers whose binary/config is missing
- STT: whisper-compatible /audio/transcriptions endpoint (cloud or
  self-hosted faster-whisper), replacing the empty dashscope stub
- config: TTS_CLOUD_URL/KEY/MODEL/VOICE, TTS_LOCAL_URL, STT_API_URL/
  KEY/MODEL; cloud keys fall back to DASHSCOPE_API_KEY
- gateway /voice/tts now returns real audio/mpeg bytes instead of a
  byte count
- tests: cloud/local/whisper round-trips via httptest, chain skip
  behavior, fallback order; kept and adapted prior service tests

backend/services/gateway/api/router.go

@@ -476,7 +476,13 @@ func (r *Router) voiceTTSHandler(w http.ResponseWriter, req *http.Request) {
 		writeJSON(w, 500, M{"error": err.Error()})
 		return
 	}
-	writeJSON(w, 200, M{"audio_size": len(audio)})
+	if len(audio) == 0 {
+		writeJSON(w, 503, M{"error": "no tts provider available"})
+		return
+	}
+	w.Header().Set("Content-Type", "audio/mpeg")
+	w.WriteHeader(200)
+	w.Write(audio)
 }
 
 // --- Admin ---