Documentation Index
Fetch the complete documentation index at: https://mintlify.com/KittenML/KittenTTS/llms.txt
Use this file to discover all available pages before exploring further.
What is Kitten TTS?
Kitten TTS is a production-ready Python library for neural text-to-speech synthesis. It runs entirely on CPU using optimized ONNX models ranging from 15M to 80M parameters, making it suitable for everything from edge devices to cloud deployments.Quickstart
Synthesize your first audio in under 5 minutes.
Available Models
Choose the right model size for your use case.
API Reference
Explore the full Python API documentation.
Voices
Learn about the 8 built-in voices and how to use them.
Key features
Ultra-lightweight
Models from 25 MB (int8) to 80 MB — designed for edge and resource-constrained environments.
CPU-only inference
ONNX-based runtime runs efficiently on any CPU. No GPU required.
8 built-in voices
Choose from Bella, Jasper, Luna, Bruno, Rosie, Hugo, Kiki, and Leo.
Adjustable speed
Control playback rate with the
speed parameter (default 1.0).Text preprocessing
Built-in pipeline expands numbers, currencies, units, ordinals, fractions, and more.
24 kHz output
High-quality audio at a standard sample rate compatible with most audio tools.
Available models
| Model | Parameters | Size | Quality |
|---|---|---|---|
kitten-tts-mini | 80M | 80 MB | Highest |
kitten-tts-micro | 40M | 41 MB | Balanced |
kitten-tts-nano | 15M | 56 MB | Fast |
kitten-tts-nano-int8 | 15M | 25 MB | Smallest |
Kitten TTS is currently in developer preview. APIs may change between releases. Check the GitHub releases for the latest updates.
System requirements
- Python: 3.8 or later
- OS: Linux, macOS, or Windows
- Hardware: Any CPU — no GPU required
- Disk: 25–80 MB per model variant
License
Kitten TTS is released under the Apache License 2.0.Community
- Discord: Join the community
- GitHub: KittenML/KittenTTS
- Website: kittenml.com
- Commercial support: Contact us