README
speech.website
Text in → speech out, running entirely in the browser: open text-to-speech models on ONNX Runtime Web (WebGPU when available, WASM otherwise). No backend; host it as static files.
Run locally
npm install
npm run dev
React + MUI frontend built with Vite; npm run build outputs
static files to dist/ (deployed to GitHub Pages by CI). The TTS libraries are
loaded at runtime from CDNs (see src/models.ts for why), and the browser
downloads model weights from HuggingFace on first use and caches them.
Contributors
18 commits
Similar repos
Languages
TypeScript
97.4%
HTML
2.6%
README
speech.website
Text in → speech out, running entirely in the browser: open text-to-speech models on ONNX Runtime Web (WebGPU when available, WASM otherwise). No backend; host it as static files.
Run locally
npm install
npm run dev
React + MUI frontend built with Vite; npm run build outputs
static files to dist/ (deployed to GitHub Pages by CI). The TTS libraries are
loaded at runtime from CDNs (see src/models.ts for why), and the browser
downloads model weights from HuggingFace on first use and caches them.
Similar repos
Contributors
18 commits
Languages
TypeScript
97.4%
HTML
2.6%