# MRAV Voice App

Real-time voice agent: Twilio Media Streams → Deepgram STT → Claude → ElevenLabs TTS → Twilio.

## Deploy to ECS

```bash
# 1. Create ECR repo (once)
aws ecr create-repository --repository-name mrav-voice-app --region us-east-1

# 2. Build & push
aws ecr get-login-password --region us-east-1 | docker login --username AWS --password-stdin <ACCOUNT_ID>.dkr.ecr.us-east-1.amazonaws.com
docker build -t mrav-voice-app .
docker tag mrav-voice-app:latest <ACCOUNT_ID>.dkr.ecr.us-east-1.amazonaws.com/mrav-voice-app:latest
docker push <ACCOUNT_ID>.dkr.ecr.us-east-1.amazonaws.com/mrav-voice-app:latest

# 3. Update task definition to use new image, then update service
aws ecs update-service --cluster mrav-voice-cluster --service mrav-voice-service --force-new-deployment --region us-east-1
```

## Twilio Setup

Point your Twilio phone number's webhook to:
```
POST https://mrav-voice-alb-865769232.us-east-1.elb.amazonaws.com/twiml
```

## Env Vars

Set these in the ECS task definition:
- `DEEPGRAM_API_KEY`
- `ANTHROPIC_API_KEY`
- `ELEVENLABS_API_KEY`
- `ELEVENLABS_VOICE_ID`
- `SYSTEM_PROMPT` (optional)