How to Make Product Photos with ChatGPT (2026): Steps, Prompts and What to Check

How to make product photos with ChatGPT: upload a clear photo of your product, then describe the new scene in one or two sentences, for example "place this bottle on a marble counter in soft morning light, square format". You get a new image in about a minute. Compare the result with the original: the model redraws everything, including labels.

ChatGPT's built-in image generation has become a popular way for small sellers to get scene photos without a photographer. It is good at mood, lighting and props. It is less reliable at keeping a product exactly as it is, which matters when the photo goes on a listing. This guide covers the steps, prompts that work, what it does well, where it goes wrong, and when a cut-out-and-composite tool is the safer choice.

How do you make product photos with ChatGPT?

Upload one sharp photo of the product and ask for a specific scene. The more concrete the request, the closer the first result.

1. Start with a good source photo. Shoot the product in daylight on a plain background, filling most of the frame, with the label facing the camera. A blurry or dark photo gives the model less to copy.
2. Open a new chat and attach the photo. Use the paperclip or plus button. Image generation is built into ChatGPT, so you don't need a separate tool.
3. Describe the scene. Name the surface, background, lighting, camera angle and aspect ratio. Add "keep the product, label and colors exactly as in my photo".
4. Check the result against the original. Zoom in on the label, logo, cap, edges and color.
5. Refine in the same chat. Ask for one change at a time: "make the background warmer", "move the product left to leave space for text".
6. Download and resize for the channel you're posting to.

For a photo that ends up on a product page, the source photo matters as much as the prompt. Our guide to taking product photos with your phone covers a simple setup.

What does ChatGPT do well for product photos?

It is strongest when the photo's job is to set a mood rather than to document the product.

• Ad and social concepts. Quick drafts of a summer scene, a holiday table or a gym bag setup, so you can see which direction works before spending on a shoot.
• Props and locations you don't have. A beach, a Scandinavian kitchen, a snowy windowsill.
• Lighting and mood. Golden hour, moody dark backdrops, bright pastel studio looks.
• Composition with space for text. It follows instructions like "leave the right third empty for a headline" reasonably well.
• Conversational editing. You refine by chatting rather than learning editing software.

What can go wrong when ChatGPT edits your product?

The model doesn't paste your product into a new scene; it draws a new picture that looks like your product. Most of the time the difference is small. Sometimes it isn't.

Simple products with no text (a plain mug, a wooden bowl) usually survive the redraw well. Products with dense labels, fine print, patterns or exact brand colors are the most likely to drift.

Is ChatGPT image generation free, and what are the limits?

Free ChatGPT accounts can generate images, but only a few per day, and OpenAI describes the free limit as variable rather than a fixed number. Paid plans raise the limit and generate faster. When you hit the cap, ChatGPT tells you when you can generate again.

ChatGPT plan names and image limits were checked in October 2026. Marketplaces change their rules, so confirm the current requirements in your seller dashboard before you upload a large batch.

OpenAI changes model names and limits often, so check the plan page or the limit message in your account rather than relying on numbers from blog posts. Before using images commercially, read OpenAI's current usage terms too.

ChatGPT product photo prompt examples

Attach your photo first, then use one of these. Each one names the surface, light, angle, format and the instruction to keep the product unchanged.

• "Place this product on a white marble kitchen counter, soft morning window light from the left, shallow depth of field, eye-level angle, square 1:1. Keep the product, label and colors exactly as in my photo."
• "Put this product on a light oak table with a linen napkin and a sprig of eucalyptus, natural daylight, 4:5 portrait. Don't change the label text."
• "Studio shot of this product on a seamless pastel peach backdrop with a round podium, soft even light, gentle shadow, 1:1."
• "Show this product on a bathroom shelf with white tiles and folded towels, spa mood, soft light, 4:5."
• "Create a 9:16 vertical scene for a story ad: this product on a picnic blanket in a sunny park, product in the lower half, empty sky in the top third for text."
• "Dark luxury look: this product on a black stone plinth, dramatic rim light, deep charcoal background, 1:1."
• "Holiday version: this product on a wooden table with warm fairy lights and wrapped gifts softly blurred behind, 4:5."
• "A hand holding this product at chest height against a blurred cafe background, natural light, 4:5. The hand should look relaxed and natural."

Follow-ups that help: "the label text changed, restore it exactly as in my photo", "make the product 20% larger in the frame", "remove the props on the left". If a detail keeps drifting after two or three tries, the model probably can't hold it; switch methods rather than prompting forever. For more prompt ideas, see our AI product photography prompts.

ChatGPT vs a cut-out-and-composite tool: which should you use?

Use ChatGPT for concepts and ads where the scene matters most; use a cut-out-and-composite tool for listing photos where the product has to be exact.

A composite tool works differently: it cuts your product out of your photo and generates only the background around it. The product pixels are your photo's pixels, so the label can't be misspelled and the shape can't change. The trade-off is that the product keeps the lighting and angle you shot it with, so the scene has to suit it.

Making listing photos with AiArter instead

AiArter (our tool) is built for the listing side of this split.

• Lifestyle scenes (8 credits). Your product is cut out of your photo and kept as-is; only the background is generated. Pick from 10 presets (Studio, Marble counter, Wooden table, Outdoors, Beach, Living room, Bathroom shelf, Pastel color, Dark luxury, Holiday) or describe your own scene in plain words in the AI background generator.
• White main image (4 credits). Pure white RGB 255,255,255, product centered at about 85% of a 2000×2000 frame, optional soft shadow.
• Channel sizes (8 credits). Instagram 4:5, Story 9:16, Pinterest 2:3, ad 1200×628 and more; AI fills the extra space instead of cropping the product.
• Bulk. Up to 50 photos per batch, one ZIP download; failed photos aren't charged.

If you want a ChatGPT-style result with a reference photo, AiArter also has an image-to-image generator running Google's Nano Banana 2 (17 credits). Like ChatGPT, it uses your product as a reference and may not match it pixel for pixel, so use it for concepts and check it the same way. For more on choosing between approaches, see AI product photo generators compared and lifestyle product photography ideas.

Frequently asked questions

Can ChatGPT make product photos from my own photo?

Yes. Upload a photo of your product and describe the scene you want, and ChatGPT generates a new image based on it. The product is redrawn rather than copied, so check labels, logos and colors against the original before using it.

Does ChatGPT change my product's label or logo?

It can. Because it generates a new image, text and logos sometimes come out misspelled or slightly reshaped, especially on busy labels. Asking it to keep the label exactly helps, but it isn't a guarantee.

Can I use ChatGPT product photos on Amazon?

Not as the main image in most cases, since Amazon expects the main image to show the actual product on pure white and accuracy matters. Generated lifestyle images can work as secondary images if they represent the product accurately. Check Amazon's current image rules first.

How many images can I make with ChatGPT for free?

Free accounts get a small daily allowance that OpenAI describes as variable, so there's no fixed number. Paid plans have higher limits. ChatGPT shows when you can generate again once you hit the cap.

What's the best prompt for ChatGPT product photos?

Name the surface, background, lighting direction, camera angle and aspect ratio, and add "keep the product, label and colors exactly as in my photo". Then refine one change at a time in the same chat.

Is ChatGPT or a background-replacement tool better for product photos?

ChatGPT is better for creative ad concepts and scenes that need new angles or hands. A tool that cuts out your product and replaces only the background is better for listing photos, because the product stays exactly as photographed.