Can I put a flat-lay clothing photo on an AI model?
10 min read
Yes. A flat lay gives an AI model generator the garment's color, print and front construction, and the AI generates the person, the pose and the way the garment falls on the body. What a flat lay can't show is drape, volume, length on the body, fit and, unless you photograph it, the back. Those parts are generated, so the result depends on how you lay the garment out and how much of the missing information you supply.
Nightjar gives that missing information somewhere to go. Save your flat lays as a Product: the front as the Main photo, a back flat lay and close-ups of print and hardware, plus a factual Description of the fabric and cut and the garment's real measurements. Nightjar treats the photos as the authority on everything visible and uses the Description and Dimensions for what the photos can't show. Then in Create choose With model, pick one of 80+ AI fashion models or let Nightjar cast one, set a Pose and crop, and generate. Compared with uploading one flat lay for each image, the Product brings the back, the details, the fabric facts and the measurements into every shot, and a saved Recipe puts every garment in your line on the same AI model, with the build you describe, in the same setup.
What can't a flat lay show, and what do I add for each part?
A flat lay records the garment's surface, not its shape on a body. On a flat surface the fabric isn't hanging from shoulders or hips, so nothing in the photo shows how it falls, how much volume the cut has or where the hem lands. Supply each missing part like this:
| Part of the garment | In a flat lay | What to add |
|---|---|---|
| Color, print and texture | Shown, when the light is even and the camera points straight down | Close-ups of the print, fabric texture, logos and trims |
| Front construction: neckline, seams, pockets, buttons, zips | Shown | Close-ups of closures and hardware |
| Sleeve length and cuffs | Shown only if the sleeves lie straight at full length | Lay the sleeves out straight at full length, never folded under or crossed |
| The back | Not shown | A back flat lay as a second Product Photo |
| Drape: how the fabric hangs and moves | Not shown | Fabric facts in the Description: weight, stiffness, stretch and how the cut falls. A hanger photo, if you have one |
| Volume and structure: shoulder pads, dropped shoulders, puff sleeves, gathers, flare, a standing collar | Flattened. Seam positions show, the shape doesn't | Describe the structure in words. For structured pieces, add a hanger or ghost-mannequin shot |
| Length on the body | Can't be judged: there's no body and no scale | The garment's real measurements under Dimensions |
| Fit, ease and size | Not shown | Nothing you can add to the photos. Keep fit information on the listing |
Write the Description for what the photos leave out. Nightjar takes visible details such as sleeve length, closures and print placement from the Product Photos, and uses the Description and Dimensions as supporting context: facts the photos can't show, such as fabric weight or stretch, are kept, and any claim the photos contradict is dropped. Measurements tell Nightjar the garment's real size, which it can use to scale the garment against the model. They inform the generated image rather than fix exactly where a hem lands, which is why length is the first thing to check afterwards.
Measure the sample you photographed and enter the measurements with units. Dimensions accepts values such as "Length 118 cm" and rejects size labels such as "EU 38", which belong on your listing.
As a hypothetical example, a seller with a linen midi dress would add the front flat lay as the Main photo, a back flat lay and a close-up of the button placket. Description: "Midweight washed linen, relaxed A-line cut with a gentle flare from waist to hem, no stretch." Dimensions: "Length 118 cm shoulder to hem, bust 50 cm measured flat." The photos carry the color, buttons and neckline. The words carry what lying flat hid.
How do I lay out and shoot a flat lay for an on-model image?
Lay the garment out to show its true outline and every part, not to style a listing. Listing flat lays often fold the sleeves, pin in the waist or add accessories, and each of those hides something the AI then has to guess.
- Steam out storage creases. Shopify's clothing photography guide says to "iron or steam everything to remove wrinkles" (Shopify). In a source photo, a fold line can look like a pleat or seam the garment doesn't have. For creases in photos you've already taken, see whether AI can fix wrinkles in clothing photos.
- Lay it in its natural shape. Square the shoulders, lay the sleeves straight at full length, straighten the hem and leave knitwear at its relaxed width rather than stretched. Nightjar takes sleeve length from the photos, so a sleeve folded under the garment gives it no length to keep.
- Shoot straight down. Center the camera over the garment and hold it parallel to the surface. A tilted camera makes the nearer end look wider and longer than it is, and that distorted outline becomes the evidence.
- Photograph one garment at a time. Leave props, accessories and other pieces out of the frame. Nightjar identifies a Product as the item its photos share and are centered on, and treats anything beside it as background, so a styled outfit grid makes the subject unclear. To combine pieces, use separate Products, as described below.
- Use a plain background that contrasts with the garment, in soft, even light. A white shirt on a white sheet loses its edges, so put it on grey or a darker surface. Avoid colored light, because the garment's color comes from the photo.
- Remove swing tags, price stickers, pins and clips. You don't need to hide the neck, size and care labels inside the garment: Nightjar is designed to keep those hidden once the garment is worn.
- Shoot the back the same way, then the close-ups: print, fabric texture, labels and hardware.
One clean front flat lay is enough to start. The back and close-ups reduce how much of the garment has to be invented.
How do I put a flat-lay garment on an AI model in Nightjar?
- Save the flat lays as a Product. Open Products, choose New product, then Manually. Add a name, add the flat lays with Upload, choose Make main on the cleanest front view, fill in Description and Dimensions, and select Create product. Creating and managing Products covers the editor. If the flat lays are already on your product page, From a link stages the name, details and photos for you to review before saving, as the product-page import guide explains.
- Open Create and select the Product. On mobile, open Generate first. Under Subject, choose With model.
- Choose the model. Select Model and pick one of the pre-built Fashion Models, Nightjar's reusable AI people, or create your own. A Fashion Model sets the person's face and hair, not their height or build. If build matters for your line, such as petite or plus-size, describe it in Custom Directions. That guides the generated person; it isn't a measurement or a fit control. If you leave Model empty, Nightjar casts a model suited to the garment and scene.
- Set the Pose and crop. Choose a Pose, then under Camera pick Full body where the Pose offers it, so the whole garment, hem included, is in frame for the length check below. Use Medium or Close-up for detail shots.
- Generate. Optionally choose a Setting and a Photography Style, then select Create Image. How to use Product Photography explains the remaining controls.
Choosing Photoshoot instead of Single shots makes four related images that vary angle, crop and orientation, so a frame can show the model from the side or back. Without a back flat lay in the Product, that view of the garment is invented.
When the setup looks right, choose Save as recipe. The Recipe keeps the Fashion Model, Pose, Camera Distance, Photography Style, setting, Custom Directions and output settings such as aspect ratio and resolution, but not the Product. You can then select the next garment and shoot it with the same Fashion Model, build direction and setup. Reusing one AI fashion model across a collection covers the full workflow.
Can I build an outfit from separate flat lays?
Yes. Save each piece as its own Product, then select all of them in one Create shot with With model. Nightjar is designed to put every selected Product on the model rather than swap one for an invented garment. When the selection isn't a full outfit, such as a shirt with no trousers, it adds neutral supporting clothes and shoes. A Photography Style sets the look of the image but doesn't restyle your pieces: the selected Products are the wardrobe.
The trade-off is reference space. A Create request uses up to five subject photos, and each selected Product's Main photo goes in first. A three-piece outfit leaves two places for back or detail views, and a five-piece outfit leaves none. Add the view that matters most, such as the back flat lay of the hero piece, from Library under Additional photos so it's included.
What should I check in the on-model image?
Check the parts the flat lay couldn't show, because those are the parts that were generated. Judge length against the body shown, not an exact position: the person's height is generated with each image. In a Full body shot, ask whether the hem lands at the knee, calf or ankle where your measured length would put it on someone of that height, and whether the sleeves end at the wrist or forearm as your flat lay shows. Check that the shoulder line, volume and flare match your garment and Description, not a generic version of that kind of garment. If any frame shows the back, compare it with your back flat lay, and don't use it if you didn't supply one. Stripes and checks have to stay continuous over folds and seams; why plaid distorts on AI models covers the fixes.
Nightjar's built-in visual review compares each output with your Product Photos and request before you see it. It looks for obvious failures, such as the model wearing a different garment, a clearly changed color, print or construction detail, or a garment worn in a way no one could wear it, and can retry at no extra Credit cost. That is protection against obvious mistakes, not a judgment of whether the length, drape or fit is true to your garment.
Treat the result as styling imagery, not evidence of fit. A flat lay contains no fit information, and research on virtual try-on found that "current VTO methods default to generating well-fitted results regardless of the garment or person size" (Karras et al., FIT, 2026). Keep the size chart and measurements on the listing.
When is a flat lay not enough?
A flat lay falls short when the garment's shape comes from its structure or the listing has to prove how it fits:
- Structured or tailored pieces shaped by padding or boning, such as blazers and corsets. A hanger photo or a ghost-mannequin photo keeps the three-dimensional shape that lying flat removes.
- Sheer fabric. Laid flat, it shows the surface beneath it, not how much skin shows through when it's worn.
- Lingerie and compression or support garments, whose fit and support are the reason someone buys them.
- Any listing that must show defensible fit or length.
For structured pieces, add or switch to a hanger or ghost-mannequin source. For the others, photograph the garment on a person.
When should I use Try On instead?
Use Try On when you already have the person image you want. Try On is an Edit Shortcut that puts "the outfit from" one image "on the person in" another, so it needs at least two images on the Edit board. The person image could be a photo you have the right to use (digital-likeness rights explains what that covers) or an earlier Nightjar on-model image whose person, pose and setting you want to keep. Edit works from the board images and your prompt, so the Product's Description and Dimensions don't go with it, and you add images from your device or Library rather than choosing a Fashion Model. For flat lays and no model, Create with a Product is the direct route. Virtual try-on versus AI fashion photography explains how the two terms differ.
Consistent and on brand AI photoshoots, optimized for conversion.
Nightjar