Making setup process better

[akshaynarisetti]

No description provided.

[ellaaicare]

🚀 Voice Mode v2 - DEPLOYED TO PRODUCTION

Status: ✅ Live and ready for iOS testing

Endpoint

wss://api.ella-ai-care.com/v2/voice?uid=<user_id>&session_id=<optional>

Health Check (Verified Working)

curl https://api.ella-ai-care.com/v2/voice/health

Response:

{
  "status": "ok",
  "version": "2.0.0",
  "endpoint": "/v2/voice",
  "config": {
    "vad_provider": "silero",
    "vad_stop_secs": 1.5,
    "stt_provider": "deepgram",
    "stt_model": "nova-2",
    "tts_provider": "openai",
    "tts_voice": "nova",
    "llm_provider": "groq",
    "llm_model": "llama-3.3-70b-versatile"
  },
  "dependencies": {
    "deepgram_key_set": true,
    "openai_key_set": true,
    "groq_key_set": true
  }
}

Production Test Results

✅ Health endpoint reachable
✅ Status is 'ok'
✅ VAD provider is Silero
✅ All API keys configured
✅ Pipeline config validation
✅ n8n config integration (with fallback)

For iOS Team

1. Connect to wss://api.ella-ai-care.com/v2/voice?uid=<uid>
2. Send raw PCM16 audio at 16kHz (same as v4/listen)
3. Receive TTS audio chunks back
4. VAD handles turn-taking automatically

Backend Monitoring

# Watch live voice sessions
journalctl -u omi-backend -f | grep "🎤\|🎙️\|✅\|❌"

Ready for end-to-end testing! 🎉