v150 · speech · privacy

Voice Memo Editor

A dictation-powered text editor using SpeechRecognition with on-device processing. Interim results appear as you speak; final results are committed to the memo. The sidebar shows the active processing path and recognition state.

Microphone access required. On Chrome 150+, on-device recognition keeps your audio entirely local — no data leaves the device.

// Request on-device recognition (Chrome 150+)
const recognition = new SpeechRecognition();
recognition.lang = 'en-US';
recognition.continuous = true;
recognition.interimResults = true;
recognition.processLocally = true; // require on-device path
await SpeechRecognition.available({
  langs: [recognition.lang],
  processLocally: true,
});

// Interim results appear as you speak
recognition.onresult = (e) => {
  let interim = '', final = '';
  for (const result of e.results) {
    if (result.isFinal) final += result[0].transcript;
    else interim += result[0].transcript;
  }
  updateEditor(final, interim);
};

recognition.start();

see also