phonon
Phonon 31M text-to-speech on WebGPU, with word timings
Get picks like this daily. Today's launch numbers, the top launches, AI and tech news, and a daily opportunity pick, straight to your inbox.
GITHUB · launched Nov 1, 2026 · 26 upvotes (top 86%) · 0 views
Niche: speech and voice ai models · 136 launches, 9 in the last 7 days · Explore this niche →
Category and tags
- Horizontal
- Model & infra
- Developer
- AI-native
- Hobby / open-source project
Visit website ↗ Source listing ↗
Could you build this?
Partial The browser UI and WebGPU audio playback are vibe-codeable, but the core value is a 31M-parameter TTS model with word-level timing. Training or tuning a model like that requires ML and speech expertise.
What it would actually take: A version that wraps an existing open TTS model can be built with ONNX Runtime Web or transformers.js and a WebGPU front end, and word timings can come from the model or a forced-alignment step. If the model itself is custom, you need speech data, a training pipeline, and ML researchers to reach comparable quality.
Competitors
Other products that read as similar to this one : 1859 launches clear the similarity bar, closest 8 shown.
Attention rank: #1528 of 1860 (itself plus its competitors, highest first, normalized so YC and Product Hunt are compared fairly).
Launched 367 days after the earliest competitor.
- Inflect TTS v2+ONNX, 9M/4M text-to-speech models running in the browser · hn · 2026-07-26 · 7 upvotes · similarity 0.70
- ito · github · 2026-10-02 · 11 upvotes · similarity 0.68
- LinnetTTS · github · 2026-11-03 · 24 upvotes · similarity 0.67
- awesome-tts-architectures · github · 2026-09-13 · 15 upvotes · similarity 0.65
- Superwhisper S1 mini · hn · 2026-10-08 · 7 upvotes · similarity 0.63
- Eloqium-TTS · github · 2026-09-11 · 6 upvotes · similarity 0.63
- ming-voice · github · 2026-10-21 · 273 upvotes · similarity 0.63
- persian_tts · github · 2026-09-16 · 125 upvotes · similarity 0.61
The niche, labels, vibe code rating and competitor comparison on this page are analysis by Nicheloom. If you use them, please credit nicheloom.com. How you may use our data.