tiny-voca: a fully open 4.5M-parameter TTS trained for under $10 of compute
A fully open 4.5M-parameter English TTS — code, weights and dataset under Apache 2.0 — trained from scratch in 7.7 hours on one RTX 4090 for $9.56; two voices, 24kHz, about 70x real-time on 4 CPU threads, running in a browser tab.
Why it matters
The extreme end of tiny-model TTS: a complete trainable speech stack for under $10 of compute, ranked #8 on the Open TTS Leaderboard.
What you could build with it
A hobbyist building a bedtime-story app could train a custom tiny voice on a single GPU for a few dollars and ship narration that runs entirely in the browser, with no speech API bill at all.
Does it hold up?
Promising but early: it ranks #8 on the Open TTS Leaderboard and runs in a browser tab, but it ships only two voices and no independent listening tests have been published yet.
Built with tiny-voca
- tiny-voca on Hugging Facehuggingface · The model page with the training recipe, cost breakdown and browser demo.
- n4ze3m/tiny-vocagithub · The project's own repo — training code, weights and dataset for the tiny TTS.
Learn more
First spotted on hf: source.
More AI models
Edge0-145B: open 145B MoE that runs on a single consumer GPUEdge0 open-sourced 145B and 120B latent-MoE models derived from Qwen3.8-Flash-Next, plus an inference framework that…models · JEV 0.74Strata: run a 125B model on a gaming PCFree, MIT-licensed local inference engine that runs Qwen3.8-Flash-Next (125B MoE, about 6B active per token) on…models · JEV 0.73Qwen4-Max: Alibaba's new flagship model goes liveAlibaba's 2.4T-parameter flagship is now generally available via API, with a 1M-token context, native vision, and…models · JEV 0.67smolmaxxing: run 16B LLMs on a 1 GB TV box or a 2013 iPadAn MIT-licensed llama.cpp patch set that runs language models up to a 16B MoE on devices with about 1 GB of usable RAM…models · JEV 0.65Blue TTS v3: a 388MB open text-to-speech model with voice cloningAn open-source 388MB text-to-speech model (MIT) with voice cloning from about 10 seconds of speech, 7 languages, 15…models · JEV 0.65Haru-4BOpen full-duplex spoken-dialogue foundation model that listens and speaks simultaneously (7.5 Hz frames, 133 ms per…models · JEV 0.63
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.