How to Photograph Clothes Flat for Perfect AI Mockups

AI mockup generation starts from your flat photo, and its quality ceiling is set there: a clean, well-lit flat lay produces a mockup that looks like a campaign image, while a wrinkled shirt on a bedsheet produces a wrinkled mockup on a bedsheet-colored model. The good news is that a great flat lay takes two minutes once you know the rules.

The seven rules

• 1. Shoot straight down. Position the camera directly above the garment’s center — angled shots distort proportions, and the AI will faithfully reproduce the distortion.
• 2. Use window light, not lamps. Soft daylight from one side gives even illumination without harsh shadows; overhead room lights create hotspots on fabric.
• 3. Kill the wrinkles. Steam or smooth the garment — wrinkles read as intentional texture and survive into the generated image.
• 4. Plain, contrasting background. A white or gray surface for dark garments, dark for light ones. Busy backgrounds compete with the garment’s silhouette.
• 5. Show the full garment with margin. Leave 10–15% of empty space around the edges so nothing is cropped — the AI needs the complete silhouette.
• 6. Lay sleeves naturally. Slightly angled away from the body reads better than pressed tight against it, and helps the AI understand sleeve construction.
• 7. Get close to the print. If the design has fine detail, take one additional close-up — it costs nothing and preserves detail fidelity.

Check yourself before you shoot

And here is what the whole flow looks like inside the app — from the flat photo you just prepared to the finished mockup:

Common failure modes and their fixes

• Mockup looks flat or lifeless → the source photo was underexposed; reshoot nearer the window.
• Print looks warped on the model → the garment was photographed at an angle; reshoot straight down.
• Colors shifted → mixed light sources (daylight + warm lamp); turn the lamps off.
• Fit looks wrong → sleeves or hem were folded under; lay the full garment out.

Treat the flat lay as the one manual step in an otherwise automated pipeline: two careful minutes there, and every generated image downstream inherits the quality.