SPEEDCASE BUILD
Stage 5 Live Transcript + Media Alignment
Streams participant audio through the SPEEDCASE server to AssemblyAI, preserves raw provider messages, normalizes partial and final turns, maps speakers to participant identities, and aligns every word with the evidentiary recording timeline.
1 · Identity and session
Create or join a deposition room
2 · Physical devices
Grant access and choose hardware
3 · Realtime transport
Connect and publish real tracks
4 · Evidence capture
Control the durable deposition record
idle
Development capture is performed by the host browser and uploaded to the local asset store. The provider-neutral recording model is ready for the production LiveKit Egress adapter.
5 · Live transcript
Stream, normalize, and align spoken testimony
checking provider
Each participant transcribes their own microphone stream through the SPEEDCASE server. The AssemblyAI API key never enters the browser, and participant identity supplies deterministic speaker attribution.
Permission preview
not publishedGrant device access to inspect the selected camera before connecting.
Live room
0 participantsConnect to LiveKit to render local and remote media.
Host-Composite-Capture-View
1280 × 720 · 30 fpsThe exact canvas shown here becomes the organic composited deposition segment.
Playback manifest
not created- No committed segments or markers.
Normalized live transcript
not created- No transcript turns yet.
Transcript annotations
0 annotations- No transcript annotations yet.
Authoritative session
No session yet.
StreamsOptions room
No media room yet.
Recording aggregate
No evidence record yet.
Transcript aggregate
No transcript yet.