v151 · Speech · Local-first workflow

Local-first workflow

A dictation surface for privacy-sensitive notes. Toggle unspoken punctuation, pick a product workflow, and see the transcript become copy-ready without shipping audio or text to a server.

dictation surface

unspokenPunctuation Workflow

formatted result

production shape

const SpeechRecognition = window.SpeechRecognition || window.webkitSpeechRecognition;
const recognition = new SpeechRecognition();
recognition.continuous = true;
recognition.interimResults = true;
recognition.unspokenPunctuation = true;

recognition.addEventListener("result", (event) => {
  const text = [...event.results].map((r) => r[0].transcript).join("");
  editor.value = text;
});

implementation reference

Need the exact API surface, compatibility boundaries, errors, lifecycle, and source links? Read the matching gendn reference ↗