N° 002 — The method issue

Three ways to answer one question.

Every fit tool online is trying to tell you what size to order. They go about it in fundamentally different ways, and the difference decides what each one can honestly promise you.

Last updated 1 September 2026

Three technologies answer “what size am I?” when you shop online: photo estimation, generative try-on, and depth capture. They differ in what they physically record — pixels, a rendered image, or distance — and that is what decides whether a tool can give you a measurement or only an impression.

	Photo estimation	Generative try-on	Depth capture
What it records	Photographs of you	A garment image composited onto your photo	Distance — the surface of your body, in three dimensions
What it produces	Estimated measurements	A picture of how a piece might look on you	A 3D model your measurements are read from
Hardware	Any phone camera	Any phone camera	iPhone Pro — the LiDAR sensor
The question it answers	Roughly what size are you	How might this look	What are your measurements, against this brand’s own chart
Where it is weakest	Scale has to be inferred; a photograph records proportion, not size	Fit is usually disclaimed — a rendered image is not a measurement	The hardware requirement. No LiDAR, no scan

Photo estimation

You photograph yourself against a wall, usually in fitted clothing, and software estimates your measurements from the outline. It runs on any phone, which is its real advantage. The hard part is scale: a photograph records proportion, not size, so real-world dimensions have to be inferred — from your stated height, a reference object in frame, or an assumption about how human bodies are proportioned.

Generative try-on

A model composites a garment onto a photo of you so you can see how a piece might look. Google retired its standalone Doppl app on 30 April 2026 and moved the same capability into Search and Shopping, where it now sits on product listings. Try-on answers how something looks. Most implementations disclaim fit explicitly, because a rendered image is not a measurement.

Depth capture

The LiDAR sensor on the iPhone Pro line records distance rather than colour — thousands of points describing the actual surface of you. Measurements are read off that surface the way a tailor reads a tape: around the chest, the waist, the hip. It is the only phone-native capture that produces three-dimensional geometry at real-world scale, which is why Anthroutfit talks about measurements rather than impressions. The cost is the hardware. No LiDAR, no scan.

What we don’t claim

A comparison page is usually where a company explains why it wins. Here is what we will not tell you, and why each one is missing on purpose.

An accuracy figure: Every number we have is measured against a calibration mannequin, not against human bodies. Publishing it now would mean asking you to trust a number we cannot yet stand behind. We validate the pipeline against physical tape measurements and publish accuracy targets before launch.
A ranking against other tools: That comparison has not been run, so we do not get to make it. What we can say is what each method records, which is the table above.
That you will stop returning things: An outcome depends on the garment, the brand, and you. Knowing your measurements before you order is the part we can actually put in your hands.
That every iPhone can do this: Capture needs the LiDAR sensor, which ships on the iPhone Pro line. That is the price of measuring instead of estimating, and we would rather say it here than at the App Store page.

Measured, not guessed.

Anthroutfit is on the App Store for iPhone Pro. Download it, create your account, and your measurements are yours in about a minute.

Get the app Questions, answered plainly