How the AI body scan works

A scale tells you one number and hides everything underneath it. Two people at the same weight can carry very different amounts of muscle, and the same person can lose fat for a month while the scale refuses to move. The body scan exists to give your plan something better to work from.

What the scan reads

You take three photos — front, side and back. A vision model estimates the things a coach would look for in a first session:

• Body-fat estimate — roughly where you sit, and which way it is moving
• Body type and fat distribution — where your body stores it first
• Posture — forward head, rounded shoulders, pelvic tilt
• Muscle visibility — how much definition is actually there to work with
• Observations and recommendations — what the model noticed, and what to do about it

Those feed the first plan: training split, calorie target and macro balance. From the second scan onwards you also get a comparison against the previous one — what improved, what slipped, and by how much.

Taking a scan that's worth comparing

The model is reading an outline, so the photo conditions matter more than the camera. Four things carry most of the quality:

1. Plain wall, even light. Overhead light casting hard shadows will read as definition that isn't there.
2. Fitted clothing. Shorts, or shorts and a sports top. Loose t-shirts hide the shape being measured.
3. Whole body in frame. Phone propped at about hip height, roughly two metres back, head to feet visible. Hands-free capture watches your pose and counts you in, so you don't have to run back and forth or get someone to help.
4. Same conditions next time. Same room, same light, same time of day, ideally before eating. Change the conditions and you measure the conditions.

What it is not

It is not a medical measurement and not a substitute for a DEXA scan or a clinician. A photo-based estimate has real error bars, and anyone claiming otherwise is selling something. What it does well is consistency: scanned the same way each time, the direction and pace of change are dependable, and that is what a training and nutrition plan actually needs.

Your photos

Scans are uploaded over an encrypted connection and kept so later scans can be compared with earlier ones. They are not used to train models and are not shared. You can delete a single scan in the app, or remove your account and everything attached to it from the Profile tab — see deleting your account and data and the privacy policy for exactly what is stored and for how long.

Common questions

How accurate is a body-fat estimate from photos?

Treat it as a trend line, not a lab result. A photo-based estimate will not match a DEXA scan, and it is not a medical measurement. What it is good at is consistency: scan under similar conditions each time and the direction of change — down, flat, up — is reliable, which is the part that actually drives your plan.

What happens to my photos?

They are uploaded over an encrypted connection, analysed to produce your result, and kept so later scans can be compared with earlier ones. They are not used to train models and are not shown to anyone else. You can delete an individual scan in the app, or delete your account and everything attached to it from the Profile tab or via our delete-account page.

How often should I scan?

Every two to four weeks. Visible composition change is slower than most people expect, and scanning daily mostly measures lighting, posture and water retention rather than progress.

Do I need a scan to use SnapFit?

No. You can set your goal, training place and dietary preferences and get a plan without one. The scan sharpens the starting point and gives you something to compare against later.