Pure Rust Browser Port: Qwen3-TTS Runs Locally Without GPU
doodlestein · x · 2026-08-13
A developer has rewritten the Qwen3-TTS-0.6B voice cloning model in pure Rust, compiling it to WebAssembly to run entirely in the browser. The setup requires no Python, PyTorch, or GPU servers, ensuring complete data privacy and deployment convenience.
Performance metrics:
- Native CLI runs at 1.4–1.6× real time
- Browser runs at 0.31–0.43× real time
- Achieves 28/28 layer verification and zero-shot voice cloning
Users download a 2GB quantized model into local cache on first use, after which they can generate speech from text or record audio for voice cloning directly in the webpage.
More from Infra
- Dev publicly asks neoclouds for discounts on a $30k GPU compute purchase — Sauers_ · 2026-10-16
- New hardware day: local LLM deployment the moment it arrives — lxfater · 2026-10-16
- Together AI's score centering fixes RL training collapse under quantized inference — nrehiew_ · 2026-10-16
- Inference is splitting prefill and decode across specialized chips, not just H100s — AccBalanced · 2026-10-16
- Rebasely launches in beta: Git hosting built for humans and coding agents — handotdev · 2026-10-16
- Drama3 TTS lands in audio.cpp: up to 5x real-time on RTX 5090, <8GB VRAM mode too — Acceptable-Cycle4645 · 2026-10-16