AI makers · 7 min read ·

AI Product Photography With Models: On-Model and In-Hand Shots From One Product Photo

For AI product photography with models, start from one photo of the product and pick a scene that adds a person. In Toybox's Product Mockup, On a model shows the item worn or held, and In hand shows it held up close. To put a garment on a specific model photo you have permission to use, Clothing Try-On combines a person photo with a clothing photo.

AI product photography with models turns a single product photo into a shot where a person wears it, carries it or holds it, without hiring a model or a studio. In Toybox AI, two tools cover this: Product Mockup adds a made-up person around your product, and Clothing Try-On puts a garment from one photo onto a person in another. Below are the shots each one makes, notes to paste in, the Amazon and Google rules for AI people and apparel, and the checks to run before you list.

AI product photography with models: shots you can make

A product mockup with a model works for anything a person holds, wears or carries. These are the common shots for small goods, accessories and clothing, with the tool that makes each one.

Product in hand mockups

A product in hand mockup tells shoppers how big the item is, useful for small goods such as lip balm, phone cases, candles, snacks and stickers. Product Mockup's In hand scene ("Held up close") builds the hand for you. Keep the hand from covering the label with a note on grip. Pick In hand, then add this in "Anything else?":

Only the hand and wrist in the frame, holding the product upright in front of a softly blurred cafe window. The thumb rests on the side edge, away from the label. Warm morning light from the right. The product stays sharp at the middle of the frame while the background falls out of focus.

A holding product mockup also works as a second listing image next to a plain studio shot, because it answers "how big is it" at a glance.

Product model mockups for bags, hats and scarves

A product model mockup puts an accessory on a person: a tote on a shoulder, a cap on a head, a scarf around a neck. Pick On a model ("Worn or held"). Framing the person from the chin or shoulders down keeps attention on the product and away from the made-up face. For a bag:

Someone in an oatmeal knit sweater carries the bag on one shoulder, framed from the chin down so the bag is the main subject. They walk along a quiet city sidewalk under soft overcast light. The bag's front panel and strap are fully visible and sharp, and the street behind is softly blurred.

For a hat, swap the setting and framing:

Worn by a person standing in a field of tall grass at golden hour, framed from the shoulders up with the face turned to the side and partly shaded by the brim. The front of the hat faces the camera. Warm, low light from behind outlines the brim.

AI product photography for fashion and jewelry

Fashion and jewelry need the most care, because buyers judge fit, drape and tiny details. For a necklace, bracelet or pair of earrings, On a model can frame the piece close up. Pick On a model and add:

Worn by a person in a plain black crew-neck top, framed from the collarbones to the chin with no eyes in the frame. The pendant sits centered on the chest, sharp and in focus. Soft studio light from the front left puts a gentle highlight on the metal. Neutral mid-gray backdrop.

Check stone counts, clasp shapes and chain links against the real piece, because small parts can come back slightly different. The jewelry product photography guide covers props, lighting and more jewelry prompts.

Clothing on a model photo you choose

For shirts, dresses and jackets, Clothing Try-On is the closer fit. It takes two photos: a "Person" (who to dress) and a "Clothing" photo (what they'll wear), and puts that garment onto that person. Use one garment per run. For a full outfit, run it once per piece and upload the previous result as the new person photo. In a group shot, use "Notes" to pick the person and the garment, for example "use the woman on the left and only the denim jacket". For printed shirts and front-and-back sets, the t-shirt mockup page covers the options.

Which tool for which shot

Shot Tool What you upload Watch for
Small product held to show size Product Mockup, In hand One product photo Fingers covering the label
Bag, hat or scarf worn outdoors Product Mockup, On a model One product photo The made-up person can change each run
Necklace or earrings close up Product Mockup, On a model One product photo Stone counts, clasps and chain links
Garment on a specific person Clothing Try-On A person photo and a clothing photo Fit and drape are estimated
Full outfit on one person Clothing Try-On, one run per garment The previous result as the new person photo Details shifting after several runs
Skincare or cosmetics in use Product Mockup, In hand or Custom One product photo Label text and claims

For skincare, the skincare product mockup page has bathroom and vanity scenes plus the FDA claim rules.

Tips for better results with models

  • Upload the product alone. Product Mockup asks for a product photo, and its upload box notes that "A plain background works best". Shoot the item by itself rather than in someone's hand.
  • Show the whole garment for try-on. Lay the shirt flat or hang it straight, with sleeves and hem visible and no folds hiding a print, so every part you sell is in the "Clothing" photo.
  • Pick a clear person photo for try-on. Choose a front-facing photo in good light, with the person's arms away from their body so the garment's shape isn't hidden.
  • Frame away from faces. "Framed from the shoulders down" or "face turned away" in your notes keeps the product as the subject.
  • Use contrast. Dress the model in a plain color that differs from the product, so a white tote isn't carried against a white shirt.
  • Keep the same notes for a collection. The same scene and framing across every color of a bag makes a grid that looks planned, even with different people.
  • Show a range of people. For adult clothing on a model, Amazon's image guide encourages showing people with different physical mobilities, including models who use wheelchairs or prostheses. You can describe the person in your notes, but Product Mockup has no setting that picks one, so check each result.

Limits to know

AI models are invented people, and marketplaces treat them that way. Plan around these limits:

  • The person doesn't exist. On a model and In hand make up the person. Amazon asks you to tag images of photorealistic AI-generated people with contains-synthetic-performer in the file's XMP metadata, and Google Merchant Center says images made with generative AI must keep metadata that marks them as AI-generated. Never caption an AI model as a customer or reviewer.
  • No repeatable model in Product Mockup. There's no setting to choose or reuse a person. For the same face across a clothing line, use Clothing Try-On with one consented person photo for every garment.
  • Fit is a preview, not a size guide. Clothing Try-On shows how a garment might look on that person, not how it fits their size. Put real measurements and a size chart in the listing.
  • Real people need consent. Toybox's Terms of Service prohibit generating images of real people without their explicit consent, so every person photo in Clothing Try-On needs that person's permission.
  • Clothing has its own photo rules. Amazon wants adult clothing in the main image shown on a standing model, and children's swimwear, leotards and undergarments photographed flat with no model in every image. Google Merchant Center wants apparel shown on people, without cropping off the head or feet in a full-body shot, but shoes and bags on their own in the main image, with model shots saved for additional images.
  • Hands and small parts. Check every finger, the thumb's grip and small hardware such as zippers and clasps. If something is off, run it again with a note about the grip or the framing.

Google's free Product Studio is a weak fit for these shots: Google says it doesn't support images with hands, and its tips say not to start from a photo of the product on a person. For a catalog of hundreds of garments, a subscription can cost less per image: as of September 2026, Photoroom's Pro plan (US$12.99 monthly, or US$7.50 a month on yearly billing) includes AI Fashion Models at 1K resolution, with 1,000 exports a month.

How to make product photos with models in Product Mockup

Start with a product photo on a plain background, saved as a JPG or PNG no bigger than 10 MB.

  1. Sign in to Product Mockup; a free account is enough to start.
  2. Add the product in the "Product photo" slot.
  3. Choose In hand for held shots, or On a model for anything worn or carried.
  4. Under "More options", set the Size (Portrait 4:5, Story 9:16, or Auto for your photo's shape), then write framing and light notes in "Anything else?".
  5. Press "Create mockup". While it runs, the result area reads "Usually about 30 seconds".
  6. Look closely at the fingers, the face, the label and the product's size next to the real thing. Pro plans can request a fix in "Want to change something?", currently 85 credits, for instance "move the thumb off the label".
  7. Save the PNG. Before an Amazon upload, add the AI-people tag if the person looks photorealistic.

For garments on a person you choose, open Clothing Try-On instead, upload the "Person" and "Clothing" photos, and tap "Try it on". Each tool currently costs 75 credits per image. The AI product photography guide covers the full image set, from the white main photo to lifestyle shots.

Frequently asked questions

Is AI product photography with models free anywhere?

Not in Toybox. Product Mockup and Clothing Try-On currently cost 75 credits per image each, about 75 cents at top-up prices, so the 20 free credits a new account gets won't pay for a single image; a $4.99 top-up of 500 credits covers six. Google's free Product Studio is a weak fit for model shots, because Google says it doesn't support images with hands.

Can I use a photo of a real model or an influencer?

Only with their explicit permission. Toybox's Terms of Service prohibit generating images of real people without their explicit consent, so use your own photos, a friend who agrees in writing, or a model who signed a release. Never use a celebrity or someone else's social media photo.

Do I have to disclose AI models in product photos on Amazon?

Yes, when the people look real and were fully generated by AI. Before uploading, use an IPTC-compatible editor to put the keyword contains-synthetic-performer in the file's dc:subject (XMP) field, and Amazon then shows a disclosure where applicable. You don't tag images that show only real people, even ones edited with AI tools.

Can the same AI model appear in all my product photos?

Product Mockup can't promise it, because it has no setting to choose or reuse a person, so the face can change from one On a model shot to the next. For one consistent model across a catalog, photograph one person who has agreed to it, then use Clothing Try-On with that same person photo for every garment.

Can I make a product in hand mockup without showing a face?

Yes. Pick the In hand scene, which shows the product held up close, and describe the framing in "Anything else?", such as "only the hand and wrist in the frame". For On a model, framing notes like "shown from the shoulders down" leave the made-up face out of the shot and the product in the middle of it.

Sources

Make it with AI Product Mockup

Keep reading