Record-first posting flow demo

Mobile-first prototype for category selection, guided talking points, live transcription, and topic coverage before auth and upload wiring are added.

Live topic coverage

1. Setup recording

Changing category resets the suggested topics list below.
Session: idle
Recorder: idle
Web Speech: not started
OpenAI: not started
Debug: off
Planned flow
  1. Visitor picks a category and refines the guided topics.
  2. Recording starts and both transcription engines begin listening.
  3. Transcript text is scanned against the active topics list.
  4. Later this flow will upload chunks, finalize a session, and require auth before publish.

Topic coverage checklist

This mirrors the current active topics list. Yellow means a phrase was heard in an interim result. Green means it was detected in a final result.

Web Speech API transcript

OpenAI Realtime transcript (WebRTC)