Open-source OmniVoice TTS server: 0.3s latency on RTX 3080, OpenAI-compatible API
Delicious-Farmer-234 · reddit · 2026-10-09
Reddit user Delicious-Farmer-234 open-sourced a TTS server built on OmniVoice with an OpenAI-compatible API that generates a sentence in about 0.3 seconds on an RTX 3080 and can clone a voice from a short reference clip.
The server is optimized for fast generation start and processes long text paragraph by paragraph. The author uses it to turn school books into audiobooks in their own voice for listening while driving. Settings are pre-tuned but adjustable (e.g., CFG guidance scale).
Free and open source: repo at hypersniper05/open-omnivoice-tts, with a demo page for previewing all voices.
More from Infra
- DGX Spark prices skyrocket as resale markups soar — natesiggard · 2026-10-09
- OpenAI bots hit 160K fetches for nonexistent URLs in a week, sparking RL-run speculation — gaganghotra_ · 2026-10-09
- Modal's LLM Engine Advisor picks engine, model and config for your inference workload — charles_irl · 2026-10-09
- LFM 2.5 5.4B seen as better laptop pick; 8B A1B lags 2.6B dense — Aggravating-Push-207 · 2026-10-09
- Architect Fi launches Liquid Inference, a router where providers bid per-prompt across 700+ models — markjeffrey · 2026-10-09
- Open-source lithos-metal hits 200+ tokens/s/user on Qwen3.8-27B with one M5 Max — JiaZhihao · 2026-10-09