AI Body Fat Estimator

Upload a photo, get your body-fat percentage in seconds. Free, no signup.

Your photo

Your sex

Privacy notice: your data is sent to Google's AI to run the analysis — we never save it or use it for training.

How the AI estimates your body fat percentage from a photo

One estimate, four independent reads: three published body-composition formulas computed from your measurements, and an AI vision pass over your photos. Each catches what the others miss.

01You add your photo and measurements

SexAgeHeightWeight

Your photo

WaistNeckHip

Only a front photo and your sex are required — the rest sharpens the estimate.

02We estimate your body fat four independent ways

US Navy

WaistNeckHeightHip ♀

US Navy %

RFM

HeightWaist

RFM %

BMI

WeightHeight

BMI %

AI vision

AI vision %

Formulas run in your browser; the AI reads what no tape measure can.

03We combine them into one weighted estimate

US Navy %RFM %BMI %AI vision %

Your body fat %

The AI weighs its visual read against the formulas — and says so when they disagree.

The formulas behind it: US Navy, RFM, and BMI

US Navy (Hodgdon & Beckett, 1984) — Circumferences in log-based equations — typically within ±3–4% of a DEXA scan.
RFM (Woolcott & Bergman, 2018) — A simple height-to-waist ratio, validated against DXA across thousands of NHANES participants.
BMI (weight vs. height) — The weakest of the three on its own — used only as broad context, never as the answer.

How accurate is a photo body fat estimate?

Honest limits: any photo-based estimate carries a few percentage points of margin — lighting, pose, and camera distance all move the needle, which is why results ship with a range and a margin rather than false precision. The population curve is fitted to NHANES 2018 US data, and none of this is medical advice. For tracking real change, consistent conditions week to week beat any single reading.

Frequently asked questions

Expect a margin of a few percentage points — a single front photo typically carries ±4–5%, and adding side/back photos plus tape measurements narrows it. That's why results ship with a range and a stated margin instead of false precision. Under consistent conditions, photo estimates are excellent for tracking direction over time.

No. Your photos are sent to Google's AI for the single analysis call and are never saved, never used for training, and never touch our servers. There's no account and no record of your upload.

It combines two independent reads. Classic formulas — US Navy, RFM, and BMI — are computed from any measurements you provide, right in your browser; the AI then analyzes your photos for muscle definition and fat distribution, weighs that visual evidence against the formula anchors, and returns one number with a range and confidence.

Because body fat distribution and essential-fat levels differ substantially between male and female physiques — the same visible leanness can sit 8–10 percentage points apart. The formulas, category ranges, and population curve are all sex-specific, so the answer changes materially.

Even lighting, fitted clothing (or shirtless), a relaxed front-on pose, and the camera at chest height a few steps away. Then switch to Advanced mode and add side and back photos plus your waist and neck measurements — each extra input tightens the estimate.

For most people the Lean-to-Average bands are the sustainable everyday zone — roughly 11–22% for men and 18–28% for women. Run a scan and you'll get the full category table for your sex with your result marked on it. Leaner isn't automatically healthier: essential fat exists for a reason, and the very low bands are competition territory, not a casual target.

Yes — free, no signup, no email. Scans are capped at 3 per month so the tool stays available to everyone, and the allowance resets automatically at the start of each month. It's built by the team behind Body Tracker, a progress-photo app for iOS and Android that tracks your transformation over time.