AI Image Prompt Library
62 prompts that turn a plain product photo into a finished marketing image — studio shots, billboard mockups, social visuals. Written for ChatGPT and Gemini, free to copy.
What makes these different: every prompt explains why it produces what it produces, and which words to change for your own product. Copying a prompt gets you one image. Understanding the structure lets you write your own.
New to image prompting? The guide to AI product photo prompts explains the structure every prompt here follows, so you can adapt them rather than only copying them.
Browse by category
Product Photography
Turn a phone snap of your product into something that looks shot in a studio.
13 prompts →Billboard & Mockups
Place your product on a Times Square screen, a roadside billboard or a shop window.
9 prompts →Social Media Visuals
Feed-ready images, carousels and thumbnails that match a brand look.
10 prompts →Food & Drink
Menu photography, café shots and beverage stills.
9 prompts →Headshots & Brand
Professional portraits, team photos and brand identity visuals.
9 prompts →Fashion & Apparel
Turn a hanger shot or flat lay into on-model, catalogue and lookbook images.
6 prompts →Portraits & People
Cinematic, film and street-style treatments for photos you already have.
6 prompts →Are ChatGPT “secret codes” for photos real?
Short answer: no. If you arrived here after seeing a reel about secret ChatGPT codes — /impactfx, /droneview, /cinematicphoto — those are not commands. ChatGPT has no slash commands for image styles. Type one on its own and nothing happens, which is why so many people conclude the trick “stopped working”.
What those creators are doing is giving a nickname to a long prompt they keep off-screen. The short name is the hook that makes the video watchable. The prompt is the part that produces the image, and it is the part that gets left out.
So this library publishes the prompts in full — three hundred words each, naming the lighting, the lens, the aperture and the framing, because that is what image models actually respond to. The short names are kept as labels so a prompt is easy to find again, not because typing them does anything.
If you came looking for a specific one: cinematic portrait, drone view and film-look prompts are in Portraits & People; splash, impact and luxury product prompts are in Product Photography; and on-model and catalogue prompts for clothing are in Fashion & Apparel.
Why this library is free
This started with a comments section. You have probably seen the videos — someone's product on a Times Square screen, a bottle floating in a frozen splash, a face on a roadside billboard. Underneath, the same question over and over: how do I make this?
The answer almost never comes. Instead: “DM me”. “Follow first, then I'll send it”. “Comment PROMPT and I'll reply”. Hundreds of people ask, a handful get an answer, and the rest follow an account hoping for something that never arrives.
It works as a growth tactic, which is exactly why it is everywhere. But a prompt is a paragraph of text. It costs nothing to give away, and holding it hostage for a follow wastes the time of everyone who genuinely wanted to learn something.
So every prompt here is on the page. No email, no follow, no DM, no account. Copy it and go. And because a prompt you do not understand is one you have to keep asking for, each one explains which words are doing the work — so the next time you want something slightly different, you can write it yourself.
How to write a prompt that works
Almost every prompt in this library follows the same order, and it is worth learning because it works for prompts we have not written yet:
subject → shot type → lighting → environment → lens → finish
Image models handle that sequence far better than a list of adjectives, because each clause narrows the one before it. “A blue bottle, beautiful, professional, 8K” gives the model nothing to act on. “A blue bottle, three-quarter view, lit by a single softbox from the left, on a concrete surface, 85mm lens, muted colour grade” describes a photograph that could actually be taken.
Two habits matter more than any keyword. Name the light — direction, quality and source — because lighting is what separates a product photo from a snapshot. And say what you do not want; models decorate by default, so “no gradients, no text, no decorative shapes” often does more than another sentence of description.
Where to run these prompts
All three of the tools below accept an image plus a text prompt, which is what these are built around. Free tiers are enough for everything here.
- ChatGPTUpload your photo, paste the prompt. Best at following long, structured instructions and keeping your product recognisable.
- GeminiStrong at photorealistic lighting and mockups. Handles the billboard and storefront prompts particularly well.
- MidjourneyBest image quality, but it does not transform an uploaded product photo the way the other two do. Use the prompts marked for Midjourney, which describe the subject instead.
Common questions
Do these work in both ChatGPT and Gemini?
Most do, and each prompt lists the tools it was written for. ChatGPT and Gemini both accept a reference image alongside the prompt, which is what these are designed around. Midjourney works differently, so its prompts describe the subject rather than referencing an attachment.
Do I need to upload my own photo?
For most of them, yes. The prompts say “the attached product” because they transform a photo you already have. A plain phone snap on a table is enough — the prompt supplies the lighting and the setting.
Can I use the results commercially?
The prompts are free to copy and use, including for client work. What you may do with the generated images depends on the terms of the AI tool you used, not on us — check the current terms before selling output.
Why does my result look different?
Image models are not deterministic, so the same prompt gives a different image each run. Run it two or three times and keep the best. Clean input matters more than clever wording — a well-lit reference photo beats any amount of prompt tuning.
Are these prompts copied from other sites?
No. Every prompt here was written for this library, along with the explanation of why it works. That is also why they are structured differently from the short command-style prompts circulating on social media.
Made an image you like?
Turn it into a shareable animated card with the wish card maker, or read the guide to AI product photo prompts for why each part of a prompt matters.