The dirtiest part of selling on Grailed for our team has always been the photo edit pass. A reseller shoots a flat-lay, opens Photoshop, lassoes the hoodie, refines the edge for the lace cuff, and fights with the magic-wand tool because the dark grey on the floor reads almost the same as the dark grey on the sleeve. Forty minutes per garment. Sometimes longer when the cuff has any kind of detail.
That whole pipeline collapses if the measurement and the cleaned-up product photo come out of the same shutter tap. TrueShape™ is the part of Size AI that handles the photo cleanup. Capture a garment, get the measurements in 0.92 seconds, and 1–2 seconds later a clean version of the photo lands in the Results screen, ready to drop into a marketplace listing.
What follows is what TrueShape does, when to use it instead of AI Photo Studio, and how to get the cleanest output from a flat-lay.
What TrueShape actually does
TrueShape is automatic background removal for the photo you just took. The captured image runs through an on-device edge-detection model that walks the same depth map the measurement engine used. The result is a clean cutout of the actual garment. Lace cuffs, fringe, distressed seams, and subtle drape are preserved at the real proportions of the piece. It's your photo, with the floor and the rest of the room subtracted out.
Two seconds of processing, give or take. The cleaned image lands on its own tab in the Results screen, next to the measurement view. Every capture generates one. No separate workflow, no quota, no upgrade prompt, no toggle to switch on.
The output is a 1:1 match for the e-commerce ghost-mannequin look Grailed and StockX buyers expect on premium listings. Drop it into the listing as the first photo. Add the measurement screenshot below it. The whole presentation reads as professional in a way a basement flat-lay does not.
TrueShape vs AI Photo Studio
Most resellers get confused the first time they see both features in the app. The two do different things, and picking the right one matters for marketplace AI-disclosure rules.
TrueShape is the real garment, with a clean background. The cloth in the image is the cloth you measured. Proportions are preserved exactly. Color, drape, and edge detail come from the actual photo. Use TrueShape as the lead listing photo on any marketplace.
AI Photo Studio generates new images from your flat-lay. A model wearing the garment, a ghost mannequin, a lifestyle scene. Those are AI-generated visualizations, not photos of the actual cloth. They work well for social posts and lifestyle imagery. They are not the photo a buyer should be using to judge fit.
When in doubt: TrueShape for the listing thumbnail and grid, AI Photo Studio for the supporting lifestyle shots. The Photo Studio guide covers the disclosure rules for AI-generated imagery in detail.
How to use TrueShape
TrueShape runs automatically with every capture. There is no toggle to flip, no setting to configure, and no separate flow to start. The three steps below are the standard capture flow with TrueShape's output added.
- Capture the garment. Follow the capture guide. LiDAR-equipped iPhone, flat surface, contrasting background, 90-degree angle. The shutter fires in 0.92 seconds and the depth map is in.
- Open the Results screen. TrueShape sits as a tab next to the measurement view. The cleaned image appears 1–2 seconds after the capture finishes.
- Export to your marketplace. One-Tap Export pushes the TrueShape photo directly into the listing draft on eBay, Poshmark, Depop, Mercari, Grailed, Etsy, Shopify, Amazon, and Online Store mode.
Tips that actually move the needle
- Contrast helps edge detection. Same rule as the measurement pass: light fabric on a dark surface, or vice versa. A black hoodie on a black sheet still gives a usable cutout, but a black hoodie on a tan hardwood floor reads cleaner.
- Fine details survive the cut. Lace cuffs, fringe trim, distressed denim hems, and the deliberate raw edge on a vintage band tee all come through. The edge model walks the depth map at high resolution, so fine geometry like individual lace eyelets stays intact.
- Wrinkles travel. A wrinkled flat-lay produces a wrinkled cutout. The background goes; the fabric texture does not. Smooth the cloth before the capture if you want a clean product shot.
- Avoid busy backgrounds. Patterned bedspreads, rugs, anything with strong texture. The edge model can resolve those, but the cutout quality drops.
Common questions
Is the TrueShape image a real photo?
Yes. It's the photo you captured, with the background removed. The cloth in the image is the actual garment. Proportions, color, and edge detail are preserved from the original capture.
Does TrueShape work on transparent fabrics like lace or chiffon?
Yes, with caveats. The edge detection performs well on lace and fringe due to their geometric edges. Sheer materials like chiffon and very light mesh may leave faint outlines of the surface beneath. A high-contrast background can mitigate most of these issues.
Can I export the TrueShape image to my marketplace?
Yes. The One-Tap Export workflow tool pushes the TrueShape photo into the listing draft on eBay, Poshmark, Depop, Mercari, Grailed, Etsy, Shopify, Amazon, and Online Store mode. Full coverage of the export pipeline lives in our workflow tools guide.
Is TrueShape included in the free trial?
Yes. Every Size AI tier includes TrueShape with no usage cap. The 7-day free trial gives you full access. No credit card required.
How is this different from AI Photo Studio?
The difference lies in the image type. TrueShape provides your actual photo, cleaned up, while Photo Studio creates new images of models or ghost mannequins based on your photo. Use TrueShape for the product shot in listings and Photo Studio for accompanying social and lifestyle content.
Related
- Capture. TrueShape runs on the same captured image. Get the capture right and the cutout follows.
- Photo Studio. Generated lifestyle images for social and mood-board use, with the disclosure rules that come with AI-generated imagery.
- Building a marketplace, retail tech stack, or fit-data layer? Talk to our team about API and SDK access.
