Kitten TTS has released three new open-source text-to-speech models with 80M, 40M, and 14M parameters, designed for efficient on-device use. The 14M model achieves state-of-the-art expressivity for its size, and all models are quantized to run on low-end hardware without GPUs. This release aims to bridge the gap between cloud and on-device TTS by providing production-ready, tiny models for voice agents and applications.
Background
Kitten TTS is an open-source project focused on developing tiny, expressive text-to-speech models optimized for on-device applications, addressing the need for efficient AI models that run on low-resource hardware.
- Source
- Hacker News (RSS)
- Published
- Mar 19, 2026 at 11:56 PM
- Score
- 7.0 / 10