Open-source 31M-param Phonon TTS runs 25 natural English voices locally on iPhone and Mac
amos_gyamfi · x · 2026-10-28
gradium open-sourced Phonon TTS inference (xn-ptts, Apache-2.0/MIT, 349 stars): a 31M-parameter text-to-speech model bringing 25 natural English voices (US/UK, 13 female, 12 male) to iPhone and Mac, fully local with no cloud round trip — built for offline assistants, accessibility tools, and interactive experiences.
The runtime is written in Rust with no PyTorch dependency and targets iOS, Android, browsers (wasm), and local services; one line of uvx ptts --lang en "..." writes a WAV. Ships with an OpenAI-compatible server, WebSocket server, CoreML, QNN, and PyO3 bindings.
Related event: Gradium open-sources Phonon, an on-device TTS model for iPhone and Mac(3 posts)→
More from Multimodal
- AI-generated Star Wars-style game demo shows a flyable desert world with TIE fighters — jasonkneen · 2026-10-29
- LiquidAI's LFM3-Audio-2B audio model trends on Hugging Face — LiquidAI · 2026-10-29
- Generating pop-art Halloween portraits with Flora and a vibe board — round · 2026-10-29
- Generating K-Pop music videos with a Claude Music Video skill, no MCP needed — AIandDesign · 2026-10-29
- Granary-v2 open speech model expands to 48 languages with open data and pipeline — rdesh26 · 2026-10-29
- ciguleva shares surreal AI artwork "Under the Hood of the Universe" — ciguleva · 2026-10-29