Skip to content

Trusted Faces

Trusted Faces is an optional, on-device feature: when MacWatcher recognizes an enrolled face, it suppresses the alert for that capture. The photo is still saved to your local log — recognition only silences the notification/alarm, it never skips the evidence.

Trusted Faces flow: a camera frame is aligned and embedded on-device, compared to enrolled faces; a match mutes the alert (the photo is still saved), no match alerts as usual.

Everything here runs on your Mac. No face data is ever uploaded.

  1. Open Settings → Trusted Faces.
  2. Turn on Recognize trusted faces.
  3. Click Add a trusted face and follow the on-screen guide — look at the camera and move your head slightly so it captures a range of angles.
  4. Save. You can enroll more than one person (for example, household members).

Enroll a good spread of angles and lighting for the most reliable recognition.

  • Threshold — how strict the match must be. Higher is stricter (fewer false matches, but it may occasionally miss you). MacWatcher ships a sensible default calibrated for the built-in model; only change it if recognition is too strict or too loose for your setup.
  • Confirmation frames — a face must match across several consecutive frames before it’s trusted, which prevents a single fluke frame from trusting a stranger.

Face recognition is 100% on-device. Enrolled face data and the model stay in the app’s local storage; nothing about your face is sent to Amrah Studio or anyone else.

See Troubleshooting → Trusted Faces.