# Piper GLaDOS Medium Voice for Home Assistant

This container provides a custom GLaDOS voice model for Home Assistant's Wyoming Piper TTS integration.

## Model Information

- **Voice**: GLaDOS (from Portal video game series)
- **Language**: English (US)
- **Quality**: Medium
- **Sample Rate**: 22.05 kHz
- **Source**: https://huggingface.co/Stoned-Code/piper-en_US-glados-medium/tree/main

## Quick Start

### Build and Run with Docker Compose

```bash
# Build and start the container
docker-compose up -d

# Check logs
docker-compose logs -f piper-glados
```

### Manual Docker Build

```bash
# Build the image
docker build -t piper-glados-medium .

# Run the container
docker run -d \
  --name piper-glados-medium \
  -p 10200:10200 \
  --restart unless-stopped \
  piper-glados-medium
```

## Home Assistant Integration

1. **Add to your Home Assistant configuration.yaml**:

```yaml
tts:
  - platform: wyoming
    host: <container-ip-or-hostname>
    port: 10200
    name: "Piper GLaDOS"
```

2. **Or use Home Assistant UI**:
   - Go to Settings → Devices & Services
   - Click "Add Integration"
   - Search for "Wyoming Protocol"
   - Enter the container's IP address and port 10200

3. **Test the voice**:
   - Go to Developer Tools → Services
   - Select `tts.speak`
   - Choose entity: `tts.piper_glados`
   - Enter test message and target media player

## Available Voice

Once integrated, you'll have access to:
- Voice ID: `en_US-glados-medium`
- Language: English (United States)
- Speaker: GLaDOS

## Container Details

- **Base Image**: `rhasspy/wyoming-piper`
- **Protocol**: Wyoming (compatible with Home Assistant 2023.5+)
- **Port**: 10200 (Wyoming protocol)
- **Model Path**: `/usr/share/piper-voices/en_US/glados/medium/`

## Troubleshooting

### Check if the service is running:
```bash
curl http://localhost:10200/api/voices
```

### View container logs:
```bash
docker logs piper-glados-medium
```

### Test TTS directly:
```bash
echo "Hello, subject. The tests are now complete." | docker exec -i piper-glados-medium piper --model /usr/share/piper-voices/en_US/glados/medium/en_US-glados-medium.onnx --output-file - | aplay
```

## Model Files

- `model/en_US_glados_medium.onnx` - Neural network model
- `model/en_US_glados_medium.json` - Model configuration and phoneme mappings