Hey Product Hunt! 👋 I built this because every "free online TTS" I tried had the same catch: your text gets uploaded to someone's server, you hit a daily quota, or you're forced into an account. So this one runs the whole speech model in your browser. It's MOSS-TTS-Nano, a 0.1B multilingual speech model from OpenMOSS, compiled to ONNX and running on ONNX Runtime Web's WebAssembly backend. Your text and the generated audio never leave your device — there is no server-side API call at all. What it does: type a sentence, pick a built-in voice (or upload a short clip to clone one), hit Speak, then download MP3 or WAV. Speed and volume controls included. No signup, no watermark. What's not great yet: the first load pulls ~760 MB of weights from Hugging Face (cached after that), and the speed slider resamples the waveform, so faster audio also sounds slightly higher-pitched — tape-style. Voice cloning on short reference clips is hit or miss. I'd love your feedback on: which languages you tested, and whether CPU inference speed feels usable on your machine. What would you add next — SSML, longer documents, or batch mode?

Text to Speech Local

Free text to speech that runs entirely in your browser

Free text to speech that runs entirely in your browser