Text to Speech (TTS)

Convert text to spoken audio locally in your browser. Powered by VITS WebAssembly models.

Home/Text to Speech (TTS)

Text to Speech (TTS)

Convert text into natural-sounding speech locally in your browser. Powered by VITS WebAssembly models.

Characters: 98All synthesis happens offline on your machine

Audio Output

Synthesized speech will appear here

Click 'Generate Spoken Audio' to begin

How it works:

  • ONNX Runtime compiles the speech model in a background Web Worker.
  • The first generation downloads the selected voice model (~15-40MB).
  • Downloaded models are stored in the browser cache (OPFS) for instant, offline generation on subsequent attempts.

About Local Web-TTS (VITS)

Traditional text-to-speech systems rely on cloud servers (like Google TTS or OpenAI Audio) to process text and return audio streams. This tool compiles and performs neural text-to-speech inference completely in your browser, maintaining full privacy and working completely offline.

Advantages

  • 100% Private: Your input text never leaves your device.
  • Zero Server Costs: Performs rendering utilizing your client CPU.
  • Offline-Capable: Caches models using the Origin Private File System (OPFS).

Technical Core

  • VITS Architecture: End-to-end Variational Inference with Adversarial Learning.
  • ONNX Runtime Web: WebAssembly backend for fast neural network execution in the browser.
  • Piper TTS: Rhasspy's Piper project weights optimized for clean, speech synthesis.