If Confidant won't generate notes and you're seeing errors about the model failing to load, it's almost always a RAM issue.
Confidant requires 8 GB of RAM minimum
Below 8 GB the note-generation model can't load reliably — the app will fail rather than run something that produces unreliable notes. 8 GB runs the Standard model; 16 GB gives Pro comfortable headroom.
If you have 8+ GB and the model still won't load
Check actual free RAM. Open Activity Monitor (Mac) or Task Manager (Windows). Look at memory pressure — if you're already at 80%+ before opening Confidant, the model has nowhere to live.
Close other heavy apps. Most common offenders:
Other AI desktop apps (ChatGPT, Claude, Cursor)
Video editors (Final Cut, Premiere, DaVinci Resolve)
Multiple Chrome windows with many tabs
VMs or Docker containers
Other transcription / dictation apps
Switch to Standard. If you chose Pro on an 8 GB machine, it may run out of memory. Open Settings › On-device AI › Models › Note generation and click Use next to Standard. The change takes effect on the next note; if Standard isn't downloaded yet, click Download now.
Restart Confidant. Sometimes the AI engine gets into a weird state and a fresh launch resolves it.
Restart your computer. If you've been up for a week, RAM may be fragmented enough that even a roomy machine struggles to allocate a contiguous block.
Standard vs Pro
Confidant ships two note-generation models:
Standard — about 3.5 GB on disk. Fast and dependable; concise notes in your clinical voice. Recommended on 8 GB machines.
Pro — about 5 GB on disk. The most detailed model: longer, more nuanced notes. Needs more memory; recommended with 16 GB.
You pick one during onboarding — the wizard marks the size that fits your computer with Recommended for this computer (on Windows, Pro is only recommended with a dedicated graphics card of 5 GB or more plus more than 8 GB of RAM). Change it any time in Settings › On-device AI › Models › Note generation: Auto picks based on your computer's memory and graphics, or choose Standard or Pro directly. Picking Pro on a lower-memory machine shows a warning that it may run out of memory.
"Failed to start the AI engine"
This error means the program that runs the model couldn't launch. Causes:
Antivirus is blocking it. Whitelist Confidant in your antivirus settings — the bundled AI engine binary is what triggers most antivirus heuristics.
Port 8080 is in use. Confidant talks to the model over local port 8080. Quit any other app using that port and restart Confidant.
The AI engine binary failed to download. During onboarding, Confidant downloads the AI engine separately from the model. If the download was interrupted, retry by quitting and relaunching the app.
"Unknown model architecture"
This means the model file and the AI engine are out of sync — usually because one was updated while the other wasn't. The fix is to clear both and let Confidant re-download paired versions on next launch.
If you hit this, contact support and we'll walk you through clearing the right files for your platform. The data folder isn't affected — your patients, sessions, and notes stay intact.