WebRTC Live Session Room — React + Python FastAPI — Three Role Interfaces + Cloud Recording

Live Session Recording: MVP Requirements

1. Roles & Access

Three Isolated Interfaces: Separate views for Moderator (full control), Respondent (sees/hears interview only), and Observer (invisible "backroom" view).

Access Management: Automated generation of three unique role-based links per session. Respondents without links are blocked.

2. Automated Cloud Recording

Server-Side Logic: Starts automatically on session launch; independent of local machines.

Persistence: Continues uninterrupted if the Moderator disconnects.

3. Participant Control

Waiting Room: Respondents hold in a landing area; Moderator must manually admit them.

In-Session Management: Moderator can admit/remove participants or mute/disable respondent video at any time without corrupting the recording.

4. Privacy & Security

Face Blur: Irreversible, automated respondent-only face blurring applied before final storage.

Role Integrity: Observers must remain strictly silent and invisible to Respondents.

5. Out of Scope

Stimulus display, timers, reconnection logic, third-party imports (Zoom/Teams), multi-respondent support, and private chat are deferred.