Pocket TTS
github.com/kyutai-labs/pocket-tts- Category
- AI Tools
- Rank
- No. 2051Tools index
Previous survey · No. 2057 ·
- Pricing
- Open Source
- Type
- TOOL
- Use case
- Models: Train & Run · Design & Media
- Interfaces
- CLI
- GitHub
- 9.7k stars
- Latest release
- v3.3.0
- Date
About
A 100M-parameter text-to-speech model from Kyutai Labs that synthesises speech entirely on CPU — about 6x faster than real time on a MacBook Air M4, with roughly 200ms to the first audio chunk on two cores. It covers English, French, German, Spanish, Italian and Portuguese, clones a voice from a reference clip, and streams output so input text can be arbitrarily long. It ships as a pip-installable CLI, a Python library and a local web server, and does not need the GPU build of PyTorch.
Why it made the leaderboard
Local voice synthesis with no GPU: a 100M-parameter model that streams speech at roughly 6x real time on two CPU cores, ~200ms to first audio, with voice cloning from a short reference clip. It makes on-device TTS practical on laptops and small boxes where the GPU-class models simply will not run.
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.