Slice 1 of the audio/video chat backend is live in production.
This is the foundation for a "talk to Mira while looking at things together" experience — voice in, voice out, with the option to share what is on your screen or in your camera. The backend now ingests both audio and video frames, fans out the right modality to the right model, and streams responses back.
What you can use today
Nothing user-facing yet — slice 1 is backend-only. Voice input on the existing chat UI continues to work as before; that flow is unchanged.
What is coming in slice 2
A new voice/video composer surface in the web app. Tap to call. Toggle camera. End. Recordings are not stored beyond the session unless you explicitly save the transcript.