Phase B1: quarantine pyannote/voiceprint off the live path, remove Upload UI

Route mic permission through a dedicated extension tab; parallelize mic+tab transcription

Capture the user's own mic as a second stream in Meet recording

Fix tab capture: always get fresh stream ID on-demand + propagate offscreen errors

Fix backend round(dict) crash + tab capture permission + consent/results UX

Update recording consent message to be friendlier

Redesign onboarding to 2 steps + add recording consent prompt

Major feature update: Google Meet recording, cumulative mirror feed, UX overhaul

Fix get_sessions 500 crash + extension panel stability