Starting a session

Steps

1. Open the patient from the sidebar
2. Click the red Start New Session button near the top of their detail page
3. The recording panel slides in from the right
4. Pick the session type — individual, couples, or group
5. Click the red Record button

The status indicator at the top of the panel turns red and shows Recording, and a red bar appears at the top of every screen so you don't lose track that a session is live.

What happens behind the scenes

When you click Record:

• Confidant opens an audio stream from your selected input device
• The first time you record, macOS or Windows will prompt for microphone permission — approve it
• Audio is captured to disk in real time as a .wav file
• A live transcription thread starts processing audio chunks as they arrive (so you'll see partial text appearing if you keep the panel open)

Nothing is sent to the cloud. The audio is on your disk, and Whisper is doing the transcription locally on your CPU/GPU.

Picking the right session type

The session type tells Confidant how many speakers to expect, which improves transcription accuracy and (eventually) speaker diarization:

• Individual — 2 speakers (you + one client)
• Couples — 3 speakers (you + two clients)
• Group — 4+ speakers

Picking the wrong type doesn't break anything — Whisper still transcribes — but it can affect speaker labeling. For now, just pick what matches the room.

What if Start New Session is grayed out or missing?

A few possibilities:

• The patient is archived — archived patients can't have new sessions. Unarchive them from the patient detail page if you want to record again.
• A session is already in progress — Confidant only allows one active session at a time. The button is replaced by a red "Session recording in progress" banner. Stop or finish the current session first.
• The model isn't ready — if you're seeing a model download or verification screen, wait for it to finish.