(none)
Use Start Mic or Simulate Stream.
This page simulates two flows: (a) Microphone capture with local buffering (no upload) just to show timing; (b) Synthetic streaming transcript using a word emitter. Real systems would push audio chunks to a transcription service (e.g. Whisper / custom). Partial hypotheses arrive first, then final segments with timestamps.
Demo intentionally keeps audio local (not transmitted). Streaming simulation contains only synthetic safe text. Production: apply de-identification, secure transport (TLS), access tokens, audit logs, optional on-device models.