Session types — individual, couples, group

When you start a session, Confidant asks you to pick a type. The choice affects how many speakers Whisper expects in the room.

The three options

• Individual — 2 expected speakers. You plus one client. Default for most therapy practices.
• Couples — 3 expected speakers. You plus two clients.
• Group — 4 expected speakers. You plus a small group.

Why it matters

Whisper, the transcription engine, performs better when it knows roughly how many distinct voices to expect. Telling it 4 speakers when there are only 2 doesn't break anything, but it can lead to over-segmentation (treating one person's voice as two). Telling it 2 when there are 4 can lead to under-segmentation.

Can I change it mid-session?

No — once a session starts, the type is locked. If you started as "individual" but a partner walks in, the recording continues fine and Whisper will still transcribe everyone, but the speaker labeling may be a bit rough. You can clean up the transcript or just regenerate the note based on what was actually captured.

What about more than 4 people?

Group is currently capped at 4 expected speakers. For larger groups (5+), the transcript will still capture everyone, but speaker diarization will be approximate. We're working on better support for larger groups in a future release.

When in doubt

Just pick the closest match. The transcript and the resulting note are not as sensitive to this choice as you might expect — Whisper is robust, and the AI generating notes works from the transcript text, not from the speaker labels directly.