Patient engagement platform

Show patients how their body is changing.
Patient privacy is assured.

Two smartphone photos return a full measurement set in ten seconds. The photos are deleted on delivery. The longitudinal record lives in your environment — the portal simply renders it, month after month, so patients can see the change a scale never shows.

See the progress view ↓ Talk to us about integration

0

Patient images stored

100+

Measurements per scan

<10s

Photo to full result

Any phone

No kiosk, scale or wearable

Images deleted on delivery · No PII retained · Mobile or web-based scans · Trained on 2M+ images

The scan

Two photos. Ten seconds. A full 3D body model.

No kiosk, no rotating platform, no wearable. A patient stands in front of any smartphone camera, front and side, and the model is reconstructed before they've put the phone down.

The scan video couldn't load Try the live demo to see the scan itself.

Input → reconstruction → 3D output

01
INPUT

Two photos, any phone

One front, one side, standing in an A-pose. Fitted clothing is best; everyday clothes work too. No special lighting or background. The scan reads pose and joints first, then solves body shape.

02
MODEL

A detailed 3D body model

A purpose-built neural network — trained on over two million images spanning body types, poses, skin tones, clothing, and environments, built on Cambridge research — fuses both views into one mesh.

03
OUTPUT

100+ measurements, handed over

Chest, waist, hips, inseam, neck, arm, shoulders, bicep, thigh, forearm and more — plus an OBJ export. Any measurement between any two points on the model is available.

Then it's gone. The photos are processed and deleted. What leaves our system is a set of numbers and a mesh — addressed to you, not held by us.

Longitudinal progress tracking

Month over month — rendered from your record, not ours.

This is the patient-facing view. Each month begins with two photos — front and side — that return a full measurement set in seconds and are deleted on delivery. The body model narrows as the numbers do, so a patient on a GLP-1 or weight-management program watches their own shape change over time. Every value shown is drawn from the record your platform holds; we supply the numbers once, then forget them. The example below is a simulated 12-week change on a sample scan, not patient data.

portal.yourprogram.com/progress

Illustrative · simulated on a sample scan

Body change over time

Sponsor-branded patient view · powered by SnapMeasureAI

Patient ID · held in your system

3D body at week 12, with the areas that got smaller since week 0 shaded blue and the week 0 body as a faint outline — Smaller About the same Larger Faint outline: week 0

Week 0Week 12

−2.6 inoff the waist in 12 weeks

Measurement	Week 0	Week 12	Change
Waist	33.9	31.3	−2.6 in
Hips	39.1	37.3	−1.8 in
Chest	39.6	39.2	−0.4 in
Thigh	22.0	22.1	+0.1 in

Every scan is the same body model in the same pose, so any two scans compare point by point: where the body changed, not just by how much. Your platform holds the history and draws this view from its own record.

0 images stored 0 measurements retained by SnapMeasureAI 0 identifiers held outside your environment Rendered client-side from the sponsor's record

Two photos in, measurements out, images gone — and nothing retained on our side after every scan.

Data architecture

The line we don't cross.

Longitudinal tracking usually means a vendor accumulating years of patient photos. Ours doesn't. The measurement engine is stateless: it reconstructs, it reports, it forgets. Everything that persists, persists inside your governance.

Patient's device

Capture

Two photos plus height, taken in a mobile or web browser. Nothing to install.

Front and side, A-pose
Encrypted in transit
Any body type, skin tone, background

SnapMeasureAI · stateless

Reconstruct & report

The mesh is solved, the measurements are computed, the result is returned — in under ten seconds.

Full 3D body mesh, 100+ measurements
Images deleted on delivery
No measurement history kept
No patient identifiers, ever

← We return · trust boundary · you retain →

POST /v1/scan → 200 OK Response shape · nulls stand in for the patient's values

{
  "scan_id":      "sm_7f3a…",          // ephemeral — invalid once delivered
  "captured_at":  "2026-09-14T14:22:09Z",
  "units":        "in",
  "subject_ref":  "<your token — we never decode it>",

  "measurements": {
    "waist_circumference":  null,   // ← your patient's value lands here
    "hip_circumference":    null,
    "chest_circumference":  null,
    "thigh_circumference":  null,
    "arm_length":           null,
    /* …95+ further sites, or your own custom set */
  },

  "mesh": { "format": "obj", "vertices": 10000, "url": "<signed, short-lived>" },

  "retention": {
    "source_images":  "deleted",
    "measurements":   "not_stored",
    "identifiers":    "none_received"
  }
}

Why it matters in a weight-management program

The scale is one number. The body is a shape.

Engagement programs live or die on whether a patient has a reason to come back next month. A visible, measured record of their own body is one of the few reasons that compounds.

Cadence

A scan for every dose

Monthly-dosed therapies create a natural rhythm: one scan per dose, an objective checkpoint that arrives with the medication. Ten seconds, at home, no clinic visit — so the checkpoint actually happens.

Signal

Change the scale can't show

Across GLP-1 trials, the share of weight lost that came from lean mass has been reported anywhere from roughly a quarter to two-fifths.¹ Circumference at trunk and limb sites shows patients where their body is changing — a complement to, not a substitute for, body-composition assessment.

Persistence

A reason to stay

In a US cohort of 125,474 patients starting GLP-1 receptor agonists, 46.5% with type 2 diabetes and 64.8% without discontinued within a year.² Progress a patient can see is a lever on the months where the number on the scale stalls.

¹ Neeland et al., Diabetes, Obesity and Metabolism (2024); reported lean-mass fractions vary widely by trial, population and method.
² Rodriguez et al., JAMA Network Open (2025), retrospective cohort, adults with overweight or obesity.

Integration

Built to drop into a program you already run.

White-labelled, API-first, and deployable wherever your compliance team needs it to sit.

Delivery

REST API

Call the endpoint from your own app or web front end. A hosted capture page can be set up on request.

Deployment

Shared or dedicated servers

Dedicated servers in the cloud region you choose, for enterprise agreements.

Branding

Fully white-label

The patient sees your program. We're happy to be invisible.

Output

100+ measurements + OBJ

Custom measurement sets defined per program — any two points on the mesh.

Cadence

You set the schedule

Monthly, weekly, per-visit. Each scan is independent — no session state to maintain.

Data handling

Custom agreements

Enterprise data-handling terms, including zero-retention attestation, on request.

Let's fill in the months.

Bring your measurement schema and we'll return a populated response against it — using your data, in your environment.

Talk to us about integration Try the live demo →

Or write to us directly: info@snapmeasureai.com