Snap Pattern - Body measurements from two photos

What it does

A tailor's tape, rebuilt in software.

Head-to-foot measurements

Neck, chest, waist, hip, shoulder, inseam, arm length and the rest - the full set a pattern needs, not just three numbers.

Calibrated, not guessed

Hold any object of known size - a sheet of A4 works - and the app scales pixels to millimetres from it. Your stated height is checked against it independently.

12 pattern blocks

Shirts, blouses, t-shirts, waistcoats, trousers and skirts, plus kaftan, dashiki, buba, agbada and iro.

Tiled PDF export

Print at home across A4 sheets, tape them together, and cut. Real seam allowances, real notches.

Confidence, stated honestly

Every measurement carries a confidence level. Where the app had to fall back to a published ratio instead of measuring depth, it says so instead of quietly guessing.

Photos are never stored

Your photographs are processed in memory and discarded the moment the numbers come back. Only the measurements are kept.

How it works

From a photo to a pattern piece.

Take two photos

One front-on, one side-on, whole body in frame, holding your reference object. Both are needed: a front photo gives width, and a circumference needs depth as well.

The engine finds your outline

Pose landmarks give the anatomical heights to measure at; the body edge is recovered from the full-colour image, with the arms subtracted so they don't fuse with the torso and inflate your chest and waist.

Widths become circumferences

Width from the front, depth from the side, combined with Ramanujan's ellipse approximation. Roughly ±3.5% when both photos read cleanly.

Draft and print

Pick a garment, set fit, sleeve, neckline and length, and export a tiled PDF you can print, tape and cut.

The catalogue

Twelve blocks. 1,847 distinct drafts.

That is not marketing arithmetic. The style options change real geometry, not labels - a sleeveless top genuinely drops the sleeve piece, and a full skirt sweeps nearly twice the hem of a straight one.

Tops

Men's shirt, women's blouse, t-shirt, waistcoat

Trousers

Men's and women's, with rise and leg options

Skirts

Straight through to full, at any length

West African

Kaftan, dashiki, buba, agbada, iro

±3.5%

Typical accuracy with both photos

Pattern blocks

Named presets

Photographs stored

Built around the photograph problem

A body photo is not something to be casual with.

So we designed the app around not keeping it.

Straight to the engine, never to an image host. The photo is uploaded directly to our own measurement service. No third-party image host, no public URL, no extra hop.
Processed in memory, then gone. The engine holds the pixels only for the life of the request and overwrites the buffers on the way out. Only the resulting numbers are stored.
No face recognition, ever. Snap Pattern does not identify anyone from a photograph and creates no biometric template.
Not used for training. Your photographs are never used to train or improve any model.
The paywall lives in the database. Measurements you have not unlocked are never sent to the app in the first place - so nothing is hidden client-side waiting to be found.

Read the full privacy policy

Two photos. Every measurement.