GitHub radar
OmniVoice — Clone any voice in over 600 languages
k2-fsa’s OmniVoice is a text-to-speech tool that allows you to upload a snippet of your voice and generate any text in that voice for 600+ languages directly from your browser without signing up!
OmniVoice is a voice cloning & text-to-speech platform developed by k2-fsa supporting more than 600 languages. Upload a sample recording of someone’s voice, write some text and get synthetic speech generated in their voice and language using the same. Built using HuggingFace Zero (GPU free tier) and deployed through Gradio, there’s no installation required. This model also has the ability to control the pace and length of the generated speech. We use a blend of Qwen3-0.6B for text encoding, OpenAI Whisper for reference transcribing and our very own OmniVoice model for synthesis, amounting to roughly 1.6B parameters in total.
Why a vibe-coder should care
With OmniVoice, you can quickly and easily translate any multilingual content or dub audio recordings into other languages while maintaining the original speaker’s voice, all at zero cost and without having to sign up!
How to install
Copy this and send it to your agent — Claude Code, Codex, any of them:
Open this space https://huggingface.co/spaces/k2-fsa/OmniVoice and explain how to use it: what to upload, how to set language and voice, what audio format it accepts.
Nothing to install — runs in the browser.
Open on Hugging Face▌ More finds