


One to eight photos, here front, side and back. The pose may differ from photo to photo, and lighting changes are fine. Height, weight and sex entered once for scale.
Send one to eight photos of a person and get their body measurements back as JSON, based on ISO 8559-1. No depth sensor or measuring tape needed.



One to eight photos, here front, side and back. The pose may differ from photo to photo, and lighting changes are fine. Height, weight and sex entered once for scale.
A parametric body model is solved against every image you send, then ISO 8559-1 girths are read off the fitted mesh.
21 girths and lengths, with left and right for the limbs, returned in one response. Per-value confidence ranges are coming.
Post one to eight photos, plus the person’s height, weight and sex, in one API call.
We find the person in each photo and fit one 3D body model to all the photos at once. Most scans finish within a minute.
Fetch the measurements as JSON with the scan id, ready for sizing, pattern making or progress tracking. Where ISO 8559-1 defines a measurement, we follow it.
Internal testing against ISO 8559-1 tape-measure ground truth. Max deviation observed across the measurement set was 2 cm.
Attach up to eight views of 10 MB each; four is the recommended set. A plain REST API with idempotent requests; every field is in the API docs.
curl -X POST https://api.menzura.ai/v1/measure \ -H "Authorization: Bearer $MENZURA_KEY" \ -F photos[][email protected] \ -F photos[][email protected] \ -F photos[][email protected] \ -F photos[]=@three_quarter.jpg \ -F height_cm=183 \ -F weight_kg=80 \ -F sex=male 202 { "id": "msr_8f2c41", "status": "queued" }
curl https://api.menzura.ai/v1/measurements/msr_8f2c41 \ -H "Authorization: Bearer $MENZURA_KEY" { "id": "msr_8f2c41", "status": "succeeded", "units": "cm", "views_used": 4, "measurements": { "chest": { "value": 114.6, "confidence": null }, "waist": { "value": 86.2, "confidence": null }, "hip": { "value": 106.9, "confidence": null }, "inseam": { "value": 76.5, "confidence": null } } }
A scan covers one person: up to eight photos in, the full set of measurements out. Failed scans aren't billed.
A test key with free scans to build and try your integration. Ask us for one.
Request a test keyExcl. VAT. Only successful scans are billed, monthly. 99.9% target uptime.
Request accessA fixed number of scans a month for a fixed price; extra scans at €1.00. On-prem or region-pinned deploys, an SLA, and mesh export on request.
Talk to usPrices in euros. Cancel any time from the dashboard. Pay per scan: paid keys stop at once, and the scans made until then are invoiced at the end of the month. Bundle: it keeps working until the end of the paid month. Scans already made aren’t refunded.
One to eight per scan. Four photos work best: front, back, left and right. More than four adds little accuracy.
No, the pose may differ from photo to photo. What has to match is the shape of the photos: the same aspect ratio and orientation, such as all portrait photos from one phone. See the photo guidelines.
Ideally underwear. Otherwise, tight-fitting clothes; loose or layered clothing makes the measurements less accurate.
No. Each photo is deleted as soon as its scan is done. The measurements are kept for 90 days so your application can fetch them for size recommendations, pattern making, progress tracking or other workflows; store what you need. See Privacy.
In our internal tests, every measurement was within 2 cm of a tape measurement taken per ISO 8559-1.
ISO 8559-1, wherever the standard defines a measurement: landmarks and definitions match it, so values drop straight into existing grading and pattern workflows. A few extras common in tailoring, such as front and back rise, follow the definitions in our API docs.
We onboard a handful of teams each week and help wire the first integration ourselves.