# Voice Models — Rush Limbaugh Show

## Active Model

**rush_limbaugh_v1.pth** — Place in this directory

This model is excluded from Git (see .gitignore).
It lives on your local machine at:

```
voice_models/rush_limbaugh_v1.pth
```

## Coqui TTS Setup (Windows 7 / Python 3.8)

```bash
pip install TTS
tts-server --model_path voice_models/rush_limbaugh_v1.pth --port 5002
```

Server starts at: http://localhost:5002/api/tts

## Pipeline Flow

```
rush:show:YYYY-MM-DD (KV)
    → /api/rush-limbaugh-podcast?action=fetch&date=YYYY-MM-DD
    → /api/rush-limbaugh-tts (POST segments)
    → /api/coqui-tts (POST text)
    → localhost:5002/api/tts
    → base64 WAV audio
    → browser playback
```

## ORI™ Integration

Each show segment is an ORI™ renormalized proof surface.
The voice model synthesizes the renormalized content
in Rush's delivery — indistinguishable from the original
except for real-time content delta.

**Authority:** Joshua Lopez — DCGP.AI  
**Patent:** USPTO 19/555,951  
**System:** ORI™ Oracle Renormalizing Intelligence