AI photo prompts for headshots, products and portraits, 26 to copy (2026)

A good AI photo prompt for a professional headshot starts from your own photo and tells the tool to keep your facial features exactly as they are. Then it sets the framing, background, light, lens (85mm for portraits), final use and what to avoid. This free AI prompt library has 26 ready-to-use prompts for ChatGPT, Gemini or Copilot, sorted by use.

Add Founderz as a preferred source

What you’ll take from this

This library is built so you can copy, paste and adapt each prompt in under a minute:

• Professional headshot prompts for LinkedIn, resumes and company websites that start from your own photo without changing your face.
• Prompts to edit existing photos, plus product, food, real estate, fashion, event, creative portrait and couple or family photos.
• The method behind them: the photography fields I use to write prompts that produce realistic photos.
• How to adapt the same prompt to ChatGPT, Gemini, Copilot, Claude and Midjourney.

If you want to use these AI tools on real projects, the AI course for creatives at Founderz covers Midjourney, Adobe Firefly and more. To generate custom prompts in a minute, use the free photography prompt generator.

ChatGPT prompts for professional photos

These are the prompts I use for professional photos with AI, sorted by the result you want. Each one is written to copy and paste into ChatGPT and works in other tools with small changes. The text in brackets is my example: replace it with your case and keep the rest. Library reviewed in October 2026.

Two rules before you start. If the prompt starts from your photo, upload a sharp, front-facing reference photo with good light: the better the reference, the less the model invents. And if the first image is not the image you want, do not rewrite everything: ask for one change at a time in the same conversation.

Professional headshot photo prompts from your own photo: LinkedIn, resume and website

This is the most searched use: getting professional profile photos from your personal photos, without a studio session. The key is the first sentence, which forces the AI to keep your face.

LinkedIn-ready headshot

Use the photo I am attaching as the only reference for my face. Keep my facial features, eye shape, nose, smile, skin tone and hairstyle exactly the same.
Turn the image into a professional profile photo for [LinkedIn]: head and shoulders, looking at the camera, [warm and confident] expression, wearing [a navy blazer and white shirt].
Background: [plain light gray studio backdrop]. Soft softbox light at 45 degrees with frontal fill. 85mm lens, shallow depth of field.
Realistic skin, polished finish without heavy retouching. 1:1 aspect ratio, high-resolution.
Avoid: plastic skin, unrealistically perfect teeth, changes to the shape of my face, hands in the frame.

Resume photo with an approachable look

Starting from my attached photo, and without changing any feature of my face, create a photo for [my resume for a finance role].
Chest-up framing, slight smile, [dark gray suit jacket with a light blue shirt].
Background: [soft off-white with a subtle gradient]. Even, soft light with no hard shadows. 85mm lens.
Natural colors and realistic skin. 4:5 vertical format.
Avoid: beauty filters, a younger or slimmer version of me, an overly formal or stiff pose.

Consistent team photos for your company website

I am going to send you photos of several people on my team, one at a time. For each photo, keep that person's facial features exactly the same.
Create the same style for everyone: head and shoulders, [plain warm beige background], soft light from the left with gentle fill, 85mm lens, same framing and same head height in every photo.
Casual business clothing that matches each person's photo. 1:1 format.
Keep this style identical for every photo in the series.
Avoid: changing anyone's face, different backgrounds between photos, plastic skin.

AI photo editing prompts that keep your face

When the photo is already good and only the light, background or clothing is off, edit instead of creating new images. Say what changes and, above all, what must not change.

Improve quality, light and color

Improve this photo without changing the person: keep the face, facial features, expression, pose and hairstyle exactly the same.
Fix the exposure, balance the white balance and add soft, flattering light on the face.
Increase sharpness slightly and remove noise. Keep the skin realistic.
Do not change the background, the clothing or the framing.

Change the background

Replace the background of this photo with [a soft-focus white brick wall]. Do not touch the person: same face, expression, hair, clothing and pose.
Match the light on the person to the new background so it does not look like a cutout.
Keep the original aspect ratio and resolution.

Change the clothing

In this photo, change only the clothing to [a charcoal gray suit with a white shirt, no tie].
Keep the face, facial features, expression, hairstyle, pose and background exactly the same.
The clothing should fit naturally, with realistic fabric folds and the same light as the rest of the photo.

Gemini prompts for professional photos

Gemini accepts your photo as a reference and edits in the same conversation, so the ChatGPT prompts also work here. These three are written for how Gemini responds: it does best when the instruction to keep your face comes first and each edit asks for a single change.

LinkedIn headshot from a selfie

Use my attached photo as a reference and keep my face identical: features, eye shape, nose, mouth, skin tone and hairstyle.
Turn the photo into a professional portrait for [LinkedIn]: head and shoulders, looking at the camera, wearing [a navy blazer and white shirt].
Background: [plain mid-gray]. Soft frontal softbox light with side fill. 85mm lens look with a slightly blurred background.
Realistic skin, no heavy retouching. Square 1:1 format.
Do not change my age, my weight or the shape of my face.

Change the background without a cutout look

Keep the person in this photo exactly as they are: face, expression, pose, clothing and hairstyle.
Change only the background to [a modern office with natural light, slightly out of focus].
Match the direction and color of the light on the face to the new background so it does not look like a cutout.
Keep the original aspect ratio.

Half-body photo for a website or speaker profile

Starting from my attached photo, and keeping my face as it is, create a half-body photo for [the speakers page of a conference].
Standing, arms relaxed, body turned slightly to the left, looking at the camera. Wearing [a gray knit sweater].
Background: [light concrete wall]. Natural side light from a window. 50mm lens look. 4:5 vertical format, natural colors.
Avoid: plastic skin, deformed hands, changes to my features.

Copilot prompts for professional photos

Microsoft Copilot also creates images from text and, depending on your version, lets you upload a photo to edit it. If you use Copilot at work, these two prompts cover the most common requests: your profile photo and a professional background for Teams video calls.

Profile photo for Teams and Outlook

Use the photo I am attaching as the reference for my face and keep my facial features and hairstyle exactly the same.
Create a professional profile photo for [Microsoft Teams and Outlook]: centered head and shoulders, looking at the camera, [friendly] expression.
Background: [plain gray-blue]. Soft, even light with no hard shadows. Square 1:1 format, with the face in the center so it works in a circular crop.
Avoid: plastic skin, beauty filters, changes to the shape of my face.

Professional background for video calls

Create a background image for [Microsoft Teams] video calls: [a modern office with a light wood bookshelf, plants and natural window light].
Slightly out of focus, as if shot with a 35mm lens at f/2.8. No people and no readable text.
Keep the center of the image clear and simpler, because that is where the person will be.
Horizontal 16:9 format, neutral and warm colors.

More AI image generation prompts: product, food, real estate, fashion and events

The same fields work for any photo and any AI image generator. These product images, food shots and portraits are reusable prompt examples that are written for ChatGPT and work in Gemini and Copilot without changes.

Prompts for e-commerce product photos

In product photography, accuracy wins: shape, color and label must match what the customer receives. Use your real product photo as a reference whenever you can.

Packshot on a white background

Product photo of [a sage green glazed ceramic mug] on a pure white, minimalist background.
Three-quarter view at eye level, soft and even studio light, a subtle shadow under the product.
100mm macro lens, everything in focus at f/11. Realistic glaze texture.
1:1 format for an online store listing, high resolution.
Avoid: reflections that hide the shape, extra objects, text or logos that are not on the product.

Lifestyle photo from your product photo

Use the attached photo of my product [facial serum in an amber glass bottle] as the reference. Keep the shape, color, label and proportions exactly the same.
Place it in [a bright bathroom with a white marble countertop, a eucalyptus sprig and a folded towel].
Morning side light from a window, 50mm lens, slightly blurred background.
4:5 format for social media.

Flat lay for a catalog

Top-down flat lay of [a specialty coffee kit: a bag of coffee, a manual grinder and two ceramic cups] on [a light oak table].
Soft, even daylight, gentle shadows, everything in focus.
Orderly composition with space between objects, natural colors.
1:1 format for a catalog or online store.

Prompts for food photography

For food, warm side light and a little steam do more than any adjective. If the photo is for a restaurant menu, the dish must look like the one that is served.

Dish for a menu or social media

Photo of [a creamy mushroom rice dish in a clay pot] for the menu of [a Mediterranean restaurant].
45-degree angle, warm side light from a window, a little steam, rustic wooden table, linen napkin.
50mm lens, shallow depth of field, natural and appetizing colors.
4:5 format.
Avoid: plastic-looking food, extra cutlery, an oversaturated look.

Top-down brunch table

Top-down flat lay of [a brunch table for two: avocado toast, poached eggs, orange juice and coffee] on [a white linen tablecloth].
Soft morning light from a window, natural colors, a few crumbs for realism.
4:5 format for social media.
Avoid: plastic-looking food, perfectly symmetrical arrangement.

Prompts for real estate and interiors

For a real listing, use AI to improve the photo, not to change what the buyer will see. If you add furniture virtually, say so in the listing. For design projects, see the AI prompts for architectural renders.

Improve a listing photo

Improve this photo of [a 270 sq ft living room with a window onto the balcony] for a real estate listing.
Correct the vertical lines, balance the exposure between the window and the interior, and use natural, neutral colors.
Do not add, remove or change furniture, walls, windows or finishes.
Keep the framing of the original photo.

Virtual staging of an empty room

Virtually furnish this empty room as [a Scandinavian-style bedroom with a light wood bed, white linen bedding and a jute rug].
Keep the walls, floor, windows, doors and dimensions of the room exactly the same.
Natural daylight from the window, 24mm wide-angle lens, straight vertical lines.
Realistic proportions for the furniture.

Prompts for fashion and editorial photos

In fashion, the garment is the subject. Describe fabric, drape and color with the same detail as the light.

Outdoor lookbook

Lookbook photo of [a sand-colored linen jacket] worn by [a model in her 30s] walking along [a Mediterranean street with white walls].
Golden hour lighting from the side, 85mm lens, slightly blurred background, natural movement.
Kodak Portra 400 look with fine grain. Full body, 4:5 format.
The jacket is the focus: show the texture of the linen and how it drapes.

Studio editorial

High fashion editorial portrait of [a male model in a black oversized coat] against a plain [deep burgundy] paper backdrop.
Hard studio lighting from a beauty dish above, defined shadows, 85mm lens. A fashion portrait for a magazine spread.
Vogue-style editorial finish, rich contrast. 4:5 format.
The coat is the focus: show the structure and texture of the wool.

Prompts for event photos

To cover a real event, edit the photos you took. Generating images makes sense to promote an event that has not happened yet, as long as you do not show real people as attendees when they were not there.

Promotional image for an event

Documentary-style photo of [an after-work networking event on a city rooftop at sunset]. People in small groups, natural conversations, smart casual clothing.
Warm light, string lights in the background, 35mm lens, natural movement.
16:9 format with space on the left for a headline.
Avoid: people looking at the camera, readable logos, deformed faces or hands.

Speaker on stage

Photo of [a speaker in her 40s in a green blazer] talking on the stage of [a technology conference], with the audience slightly out of focus in the foreground.
Warm stage light, 135mm telephoto lens, natural gesture with one hand.
16:9 format for the event recap.
Avoid: readable slides, deformed hands, an exaggerated pose.

Prompts for creative portraits

Here you can take more risks with light and finish. A named lighting setup, like Rembrandt lighting, gives the model a very specific reference.

Fine art black and white from your photo

Starting from my attached photo, and keeping my facial features exactly the same, create a fine art black and white portrait.
Close-up, Rembrandt lighting from the left, dark background, [black turtleneck].
85mm lens, deep blacks, Ilford HP5 film grain. 4:5 format.
Avoid: plastic skin, changes to my features, an illustration look.

Night portrait with neon

Portrait of [a musician in a leather jacket] in [a wet alley at night]. Pink and blue neon light from the side, reflections on the wet ground.
50mm lens at f/1.4, strong bokeh, cinematic look, fine grain.
4:5 format.
Avoid: plastic skin, unreadable or invented signs.

Prompts for couple and family photos

With several people, the model tends to mix features. Upload a clear photo of each person and say which photo belongs to whom. Only use photos of people who have given you permission.

Couple photo from two photos

I am attaching two photos: the first is [my partner] and the second is me. Keep each person's facial features exactly the same.
Create a natural couple photo [walking on a beach at sunset], holding hands, relaxed expressions.
Warm golden-hour light, 50mm lens, slightly blurred background. 4:5 format.
Avoid: mixing the two faces, extra fingers, plastic skin.

Lifestyle family photo

Using the attached photos of each family member, and keeping each person's facial features exactly the same, create a lifestyle family photo [in the living room on a Sunday morning, playing a board game on the rug].
Soft window light, warm natural colors, 35mm lens, spontaneous moment.
3:2 horizontal format.
Avoid: mixing faces, deformed hands, an artificial pose.

How I built these prompts for realistic AI photos

I have been writing prompts since 2020 and 2021, before ChatGPT existed: first for ads, website content, guides and FAQs, then for the processes of entire teams. Less than four months after ChatGPT launched, that work was already generating sales, not just leads. This page applies what I learned there to photography: AI does not guess what you do not give it.

I built the free AI photography prompt generator because I tried the best-known ones and none were specific enough for each use, or taught you anything while you used it. I wanted something free that gives you a useful photo quickly and teaches you to think like a photographer, so you end up writing your own prompts.

These are the seven fields I always use, in this order:

Image prompting is a skill you can learn: effective prompts follow the same structure every time, so each one becomes a reusable prompt you can iterate and refine. Three decisions explain most results. The first is the reference: in a headshot, the sentence “keep my facial features exactly the same” goes first, because it is the instruction that breaks most often. The second is named light: “clamshell” or “window side light” gives the model a concrete setup, while “good lighting” tells it nothing. The third is the avoid list, which appears at the end of every prompt in the library.

Two lessons I learned the hard way. First: asking for a photo as if you were chatting (“make me a professional photo”) gives you the most generic image possible. The second comes from 1990, when I coded teletext pages on BBC computers at school: WYSIWYG and RIRO. What you describe is what you get, and if rubbish goes in, rubbish comes out.

How the prompt changes in ChatGPT, Gemini, Copilot, Claude and Midjourney

• ChatGPT: understands full sentences, lets you upload photos and edit the image in the same conversation. Write the aspect ratio in words and ask for one change at a time. This library works best here as is.
• Gemini: also accepts reference photos and conversational editing, so the same prompts work without changes. Repeat the instruction to keep your features if you make several edits in a row.
• Copilot: creates images from the chat and, in versions that allow it, edits a photo you upload. It works well with full sentences; ask for the aspect ratio in words.
• Claude: does not generate images. Use it to write or improve the prompt (ask it to organize it by fields and add the avoid list), then paste it into ChatGPT, Gemini or Copilot.
• Midjourney: prefers short visual phrases separated by commas, and takes settings as parameters at the end: –ar for aspect ratio, –stylize for how stylized the image is and –no for what to avoid.

This is the LinkedIn prompt adapted to Midjourney:

professional LinkedIn headshot of [a woman in her 30s], [navy blazer and white shirt], confident friendly expression, plain light gray studio background, soft softbox lighting, 85mm lens, shallow depth of field, natural skin texture, photorealistic --ar 1:1 --no plastic skin, illustration, extra fingers

Why does your AI image look fake, and how do you make it photorealistic?

When a generated image looks like a painting or an illustration, the prompt almost never has photographic information. Without a lens, directional light and a finish, the model falls back to its default style: perfect skin, flat light and colors that are too clean.

• Ask for real imperfection: “natural skin texture”, “visible pores”, “subtle fine grain”. Skin without texture is the clearest sign that AI made the image.
• Give the light a direction: “soft window side light” creates believable shadows; “well lit” produces flat light with no volume.
• Name the camera: lens, depth of field and film give photographic results and photorealistic scenes.
• Use one style only: mixing “photorealistic” with “3D render” or “illustration” confuses the model.

Negative prompts: what to exclude from your images

Negative prompts tell the model what you do not want. In conversational models, one line starting with “Avoid:” is enough; in Midjourney you write them with the –no parameter. This is the base list I copy almost every time:

Avoid: plastic skin or heavy retouching, extra fingers or hands, deformed eyes or teeth, unreadable text or invented logos, over-saturated colors, illustration or digital painting look, changes to the facial features of the reference.

Usage rights and privacy for AI-made photos

AI image generation saves hours for creative professionals and for anyone who needs a profile photo, but it does not replace your judgment. Before you use an AI image generator for something you will publish, or add it to your workflow, check three things:

• Usage rights: each platform has different terms for the commercial use of images made with AI. Read them instead of assuming.
• Real people: only upload photos of yourself or of people who have given you permission, and do not present a scene that never happened as real.
• Privacy: check in each tool’s settings whether your images can be used as training data, and do not upload confidential material.

Frequently asked questions about AI photo prompts

These are the questions people ask most often about AI photo prompts for headshots and products.

What is the best ChatGPT prompt for a professional headshot?

One that starts from your photo and defines everything a photographer would: “Use the attached photo as the only reference for my face and keep my facial features exactly the same. Create a professional LinkedIn profile photo, head and shoulders, plain gray background, soft softbox light, 85mm lens, shallow depth of field, natural skin texture, 1:1 aspect ratio.” You have the full version and variants for resumes and websites in the library above.

How do I stop AI from changing my face when I edit a photo?

Put the instruction first and be explicit: “keep my face, facial features, expression and hairstyle exactly the same; change only [the background]”. Ask for one change per message and, if the face drifts away from the original after several edits, upload the original photo again and start over.

Can I make professional photos with ChatGPT for free?

Yes. The free ChatGPT plan lets you generate and edit images, with lower usage limits than paid plans. The same prompts work on both. Limits change often, so check the current ones in your account.

Which lens should I put in a professional portrait prompt?

85mm is the standard for portraits: it flatters features and separates the person from the background. 50mm gives a more natural look and 35mm includes more of the setting, which is useful for team photos in the office. Pair the lens with the depth of field you want, for example “f/1.8, blurred background”.

Can I make professional photos with Claude?

Not directly: Claude does not generate images. It is useful for preparing the prompt. Give it your idea and ask it to organize it into photo type, subject, background, light, lens, use and avoid list. Then paste the result into ChatGPT, Gemini or Copilot with your photo.

What prompt should I use in Copilot for a professional profile photo?

The same structure: upload your photo if your version of Copilot allows it, ask it to keep your features and define framing, background, light and a square format. You have the full prompt for Teams and Outlook in the Copilot section.

Can I use an AI-made profile photo on LinkedIn or my resume?

Yes, as long as it still looks like you. Anyone who sees you later on a video call or in an interview expects to recognize you, so avoid making yourself look younger, slimmer or retouched. Use AI to improve light, background and clothing, not to change your face.

Why do my AI photos look fake or like paintings?

Almost always for three reasons: skin that is too smooth, light with no clear source and an undefined style. Ask for “photorealistic photography”, name the light (for example, window side light), add “natural skin texture” and exclude “plastic look, illustration, digital painting”.

Your next step with AI photo prompts

Pick one of the prompts in the library, change what is in brackets and run your first test. Then change a single field, such as the light or the lens, and compare. Using prompts like this, one variable at a time, teaches you what each field controls much faster than reading theory.

If you want to apply these tools to real projects with professional judgment, the AI for photographers program at Founderz is the natural next step.