Starter

Voice app starter (browser SR+TTS)

A zero-dependency browser voice application using the Web Speech API. Speech recognition and text-to-speech in under 200 lines.

Real-time transcription via SpeechRecognition. Continuous mode with interim results.

Browser-native synthesis via SpeechSynthesis. Voice selection and rate control.

No npm packages. Pure browser APIs. Works offline after first load.

Web Speech API is supported in Chromium browsers. Safari and Firefox have limited TTS support.

Quick start

git clone https://github.com/meridian/voice-starter
cd voice-starter
npx serve .

Open index.html in Chrome. Click the mic button and speak. The response is spoken back via TTS.