Note generation runs on your CPU/GPU and shares your machine's RAM with everything else. If it's taking 5+ minutes for what used to take 1, here's what to check.
Quick wins (try first)
Close other heavy apps. Anything using significant RAM or CPU competes with the note model. Common culprits: video calls (Zoom, Teams), other AI apps (ChatGPT, Claude desktop), video editors, multiple Chrome windows with many tabs.
Restart Confidant. If the model has been running for a long session and your RAM is fragmented, restarting Confidant gives you a clean slate.
Check available RAM. Open Activity Monitor (Mac) or Task Manager (Windows) → Memory. If you're at 90%+ memory pressure, the model is probably swapping to disk, which is dramatically slower.
Expected timings
Generation time depends on session length, your model tier, and your machine.
Session length | Pro on Apple Silicon | Lite on Apple Silicon |
Under 30 min | 30–90 sec | 20–60 sec |
30–60 min | 1–3 min | 1–2 min |
60–90 min | 2–4 min | 1–3 min |
90+ min | 3–6 min | 2–4 min |
Intel Macs and Windows PCs run roughly 1.5–3× slower than the figures above. Very long sessions (over 90 min) take an extra moment up front while Confidant processes the transcript in pieces — the progress label tells you which phase you're in.
Don't know which tier you're on? Open Settings → Notes → Note Generation AI Model to see the active model.
Slower than expected
If you're regularly hitting times much longer than the table above:
Check your tier. Lite is faster than Pro on the same machine. If you're on Pro and want faster generation, you can override to Lite in Settings → Notes → Note Generation AI Model — notes will be a bit shorter and tighter, but speed up noticeably.
Try the same session again. First generation after launching Confidant is slower because the model has to load into memory. Subsequent generations are faster.
Confirm hardware acceleration. On Apple Silicon Macs, Confidant uses the Mac's GPU for AI work. If something has gone wrong, generation falls back to CPU — much slower. A restart usually fixes it.
Stuck mid-generation
If a generation stops streaming text mid-sentence and just hangs:
Wait at least 90 seconds. Confidant aborts after 90 seconds of silence and shows a clear error message — give it a moment.
If it just stops without an error, the transcript may be unusually long. Confidant routes very long transcripts through a multi-step pipeline automatically; if you're seeing this on a shorter transcript, contact support.
When it's the AI, not your machine
Once in a while a generation comes back blank or cuts off in the middle of a section. This isn't a hardware problem — the AI just took a bad path on that particular run. The fix is the same: click Regenerate. The next attempt picks a different path and usually works.
If three regenerations fail in a row on the same session, that's worth reporting — it's a real bug, not a flake.