Local text-to-speech with multiple engine support.
pip install narrate[kokoro] # Kokoro 82M — fast, runs on CPU
pip install narrate[chatterbox] # Chatterbox 500M — high quality, emotion control
pip install narrate[all] # Both engines + audio playbackfrom narrate import narrate
# Uses the best available engine automatically
narrate("Hello world")
# Choose engine, voice, speed
narrate("Hello world", engine="kokoro", voice="af_heart", speed=1.2)
# Chatterbox with emotion control (0.0 = neutral, 1.0 = expressive)
narrate("Hello world", engine="chatterbox", emotion=0.8)
# Save to file without playing
narrate("Hello world", output="hello.wav", play=False)| Engine | Params | Speed | Quality | Features |
|---|---|---|---|---|
| Kokoro | 82M | Fast (CPU) | Good | 9 voices, speed control |
| Chatterbox | 500M | Moderate | Excellent | Emotion control, voice cloning |
MIT