An AI photo prompt is a written instruction that tells an AI tool what picture to create or how to change a photo you upload. This beginner guide shows you a simple five-part formula, the lighting and camera words that matter, and five free copy-paste prompts you can try today in ChatGPT, Gemini, Midjourney or Stable Diffusion.
Table of Contents
- What Is an AI Photo Prompt?
- Two Types of Prompts: Create vs Edit
- The 5-Part Prompt Formula
- Weak Prompt vs Strong Prompt
- Lighting Words That Change Everything
- Camera and Lens Words Explained
- The Protect Clause: Keep Your Real Face
- 5 Beginner Prompts You Can Copy Now
- How to Use a Prompt, Step by Step
- Which AI Tool Should You Use?
- How to Improve a Result Without Starting Over
- Common Beginner Mistakes
- Frequently Asked Questions
- Final Thoughts
Key Takeaways
- A good prompt names five things: the subject, the setting, the light, the camera look and the mood.
- Specific words beat long sentences. "Soft window light from the left" is better than "nice lighting".
- When you edit your own photo, always add a clause that protects your face and identity.
- Change one thing at a time and re-run. Rewriting everything hides what actually worked.
What Is an AI Photo Prompt?
A prompt is the text you give an AI image tool. The tool reads it and produces a picture, or edits a picture you uploaded. The quality of the result depends far more on your wording than on which tool you pick.
Think of it like briefing a photographer who cannot see your mind. If you say "take a nice photo of me", you get a random guess. If you say "outdoor portrait, golden evening light, blurred mountains behind me, natural smile", you get something close to what you pictured. A prompt is that brief, written down.
Good news for beginners: you do not need technical skill or special software. You need a clear structure, a small vocabulary of useful words, and the habit of testing one change at a time. This page gives you all three. When you are ready for ready-made options, browse our full AI photo prompt library.
Two Types of Prompts: Create vs Edit
Before writing anything, decide which job you are doing. They need different prompts.
Create prompts (text to image)
You describe a scene from nothing and the AI draws it. Every detail must be in your text: who, where, what light, what camera look. These are used for concept images, backgrounds, products and fictional characters.
Edit prompts (photo to photo)
You upload your own picture and describe the change: retouch the skin, swap the background, fix the light, restore an old photo. Here the photo already holds most of the information, so your prompt is mostly about what to change and what to leave alone. If you want to practise this type, our photo editing prompts collection is built around it.
Tip: Beginners get the fastest wins with edit prompts, because the AI starts from a real photo and has less room to go wrong.
The 5-Part Prompt Formula
Almost every strong photo prompt can be built from five parts. Write them in this order and you will rarely get stuck.
| Part | What it answers | Example words |
|---|---|---|
| 1. Subject | Who or what is in the picture, and what are they doing? | a man in his forties, navy jacket, relaxed smile, looking at the camera |
| 2. Setting | Where is it, and what is behind the subject? | mountain village street, wooden shopfronts, blurred background |
| 3. Light | Where does the light come from and how does it feel? | golden hour sun from camera right, soft rim light on the shoulder |
| 4. Camera | Lens, distance, angle and depth of field | 85mm lens, f/1.8, eye-level, head-and-shoulders crop |
| 5. Mood and quality | Colour feel, style and finishing details | warm tones, cinematic, natural skin texture, sharp focus |
Add an optional sixth part when you need it: a short list of things to avoid, such as "no extra fingers, no warped text, no plastic skin". Many tools handle this better as a separate "negative prompt" field, but writing it at the end of the prompt also works in chat tools.
Here is the formula assembled into one sentence:
Formula Example
All five parts in one line
A man in his forties wearing a navy jacket, relaxed smile, looking at the camera [subject], standing on a mountain village street with blurred wooden shopfronts behind him [setting], warm golden-hour sunlight from camera right with a soft rim light on his shoulder [light], 85mm lens, f/1.8, eye-level, head-and-shoulders crop [camera], warm tones, natural skin texture, sharp focus on the eyes [mood and quality].
Weak Prompt vs Strong Prompt
The easiest way to see the formula working is to compare a vague request with a structured one.
| Weak prompt | Strong prompt | What changed |
|---|---|---|
| "A girl in a park" | "A young woman in a yellow raincoat walking through a green park after rain, soft overcast light, 50mm lens, shallow depth of field, calm mood" | Added clothing, action, weather, light and lens |
| "Make my photo better" | "Even out the skin tone and remove small blemishes but keep pores, brighten the eyes slightly, keep the background and outfit unchanged" | Said exactly what to change and what to protect |
| "Cool city photo at night" | "Rainy night street, wet road reflecting neon signs, blurred headlights, cyan rim light, 35mm f/1.4, cinematic" | Replaced "cool" with concrete visual details |
The pattern is always the same: vague adjectives like "nice", "cool" or "beautiful" mean different things to different people, but "soft overcast light" and "35mm f/1.4" mean something precise.
Lighting Words That Change Everything
Light is the single biggest factor in how professional a photo looks. If you only learn one vocabulary set, learn this one. For a deeper set of ideas, see our golden hour prompts and studio photography prompts.
| Lighting term | What it looks like | Good for |
|---|---|---|
| Golden hour | Warm, low sun with long soft shadows | Outdoor portraits, travel photos |
| Soft window light | Gentle directional light from one side | Indoor portraits, natural look |
| Overcast daylight | Even, shadowless, slightly cool | Clean skin tones, fashion |
| Rim light / backlight | Bright glowing edge around hair and shoulders | Separating subject from background |
| Rembrandt lighting | One side lit, small triangle of light on the shadow cheek | Dramatic studio portraits |
| Softbox / beauty dish | Large soft studio light, flattering on skin | Headshots, profile photos |
| Neon / practical light | Coloured light from signs and lamps in the scene | Night city, moody scenes |
Always add the direction of the light ("from camera left", "from behind"). It stops the AI from lighting the subject one way and the background another, which is the main reason an edit looks pasted on.
Camera and Lens Words Explained
You do not need to be a photographer. These few terms are enough to control how the picture feels.
- Lens length (24mm, 35mm, 50mm, 85mm). Wide lenses (24 to 35mm) show more of the surroundings. Longer lenses (85mm and up) compress the background and are flattering for faces.
- Aperture (f/1.4 to f/8). A low number such as f/1.8 gives a softly blurred background. A higher number such as f/8 keeps everything sharp.
- Camera angle. "Eye-level" feels natural, "low angle" makes the subject look strong, "high angle" feels small and soft.
- Framing. "Head-and-shoulders", "waist-up" and "full body" tell the tool how much of the person to show.
- Depth of field. "Shallow depth of field" blurs the background. "Deep focus" keeps both near and far objects sharp.
These words describe how a real camera behaves, and AI models have learned from millions of real photos, so they respond well to them. A simple rule: 85mm f/1.8 for faces, 35mm f/2 when the place matters too.
The Protect Clause: Keep Your Real Face
When you upload your own photo, the biggest risk is that the AI quietly changes your face. It may smooth it, reshape it or blend it with a more generic face. The fix is one extra sentence, which we call the protect clause.
Protect Clause
Paste at the end of any edit prompt
Preserve the exact facial identity, bone structure, eye shape, lip shape, jawline and natural skin tone of the person in the photo. Do not beautify, slim, reshape or blend the face with any other face. Keep natural pores and fine lines.
Add it to every edit prompt. It does not guarantee a perfect result, but it removes the most common failure. For artistic styles such as painting or sketch effects, expect a small amount of drift. That is normal.
A related habit: only upload photos of yourself, or photos of people who have given you permission. Read more in our Terms & Usage Policy.
5 Beginner Prompts You Can Copy Now
Each prompt below follows the five-part formula. Copy one, replace the words in the example with your own details, and test it in your tool of choice.
1. Natural Outdoor Portrait (Create)
A simple text-to-image prompt. Use it to learn how the five parts work together without uploading anything.
Prompt 01
Best for: ChatGPT · Gemini · Midjourney
A candid outdoor portrait of a woman in her thirties wearing a cream knit sweater, gentle natural smile, looking slightly off camera, standing in a quiet green valley with blurred trees and soft hills behind her. Warm late-afternoon sunlight from camera left with a soft rim light on her hair. 85mm lens, f/1.8, eye-level, head-and-shoulders crop, shallow depth of field. Warm natural colours, realistic skin texture with visible pores, sharp focus on the eyes. Ultra-realistic, 8K. Free from distorted hands, extra fingers, warped text and plastic skin. Add a small watermark text reading 'www.PromptsForPhotos.Com' in the bottom-right corner, small clean sans-serif font, pure white, 50% opacity, unobtrusive, not overlapping the subject.
2. Natural Skin Cleanup (Edit)
Your first edit prompt. It fixes small flaws and keeps real skin texture, which is what separates a good retouch from a plastic filter.
Prompt 02
Best for: Gemini · ChatGPT
Retouch the uploaded portrait with a natural finish. Remove temporary blemishes and localised redness, even out patchy tone, and keep visible pores, fine lines and real skin texture fully intact. Do not smooth, blur or airbrush. Keep the original outfit, hair and background unchanged and match the existing light direction. Preserve the exact facial identity, bone structure, eye shape, lip shape, jawline and natural skin tone of the person in the photo; do not beautify, slim, reshape or blend the face with any other face. Ultra-realistic, 8K, sharp focus, natural skin texture. Free from distorted hands, extra limbs, warped text and plastic skin. Add a small watermark text reading 'www.PromptsForPhotos.Com' in the bottom-right corner, small clean sans-serif font, pure white, 50% opacity, unobtrusive, not overlapping the subject.
3. Golden Hour Background Swap (Edit)
Moves your subject into warm evening light. It also tells the AI to relight you, so the result does not look cut out.
Prompt 03
Best for: Gemini · ChatGPT
Replace the background of the uploaded photo with a quiet mountain road at golden hour: warm amber sunlight, distant snowy peaks softly blurred, long gentle shadows. Keep the subject's pose, outfit and framing exactly as they are, then relight the subject with warm low sun from camera right and a soft golden rim light on the shoulder so they sit naturally in the scene. Preserve the exact facial identity, bone structure, features and skin tone; do not beautify or blend the face. 50mm f/1.8 look, creamy background blur, eye-level, waist-up crop. Ultra-realistic, 8K, sharp focus, natural skin texture. Free from distorted hands, extra limbs, warped text and plastic skin. Add a small watermark text reading 'www.PromptsForPhotos.Com' in the bottom-right corner, small clean sans-serif font, pure white, 50% opacity, unobtrusive, not overlapping the subject.
4. Clean Studio Headshot (Edit)
A reliable profile photo look for CVs, LinkedIn and company pages. See more in our headshot prompts.
Prompt 04
Best for: Gemini · ChatGPT
Turn the uploaded photo into a clean professional headshot. Set the background to a soft dark-grey studio gradient. Light with a large softbox at camera left, 40 degrees, plus gentle fill from the right, with a small catchlight in each eye. Tidy the collar and shoulders, reduce oily shine while keeping natural skin sheen, and keep the hairstyle and expression unchanged. Preserve the exact facial identity, bone structure, features, skin tone and expression; do not beautify, slim or blend the face. 85mm f/4 look, eye-level, head-and-shoulders crop, 4:5 vertical. Ultra-realistic, 8K, sharp focus on the eyes, natural skin texture. Free from distorted hands, extra limbs, warped text and plastic skin. Add a small watermark text reading 'www.PromptsForPhotos.Com' in the bottom-right corner, small clean sans-serif font, pure white, 50% opacity, unobtrusive, not overlapping the subject.
5. Cinematic Rainy Night Street (Create)
A mood-heavy create prompt that shows how light and colour words carry a whole image.
Prompt 05
Best for: Midjourney · ChatGPT · Gemini
A cinematic portrait of a man in a dark coat standing on a rainy night street, wet asphalt reflecting magenta and cyan neon signs, blurred headlights, fine rain visible against the light. Cyan rim light from behind left, soft magenta accent on the right cheek, low ambient light. 35mm lens, f/1.4, slight low angle, waist-up crop, shallow depth of field. Moody teal and magenta colour grade, realistic skin texture, sharp focus on the face. Ultra-realistic, 8K. Free from distorted hands, extra fingers, warped text and plastic skin. Add a small watermark text reading 'www.PromptsForPhotos.Com' in the bottom-right corner, small clean sans-serif font, pure white, 50% opacity, unobtrusive, not overlapping the subject.
How to Use a Prompt, Step by Step
Copy the prompt ➡️ open your AI tool ➡️ upload your photo if it is an edit prompt ➡️ paste ➡️ generate ➡️ adjust one thing and run again.
- Choose the job. Create (no upload) or edit (upload your own photo).
- Use the sharpest source. For edits, upload the original file, not a screenshot. Compressed copies give the AI less detail to work with.
- Upload first, then paste. In ChatGPT and Gemini, attach the photo before you send the text.
- Replace the example details. Change clothing, colours, place and mood to match what you want.
- Generate two or three versions. Pick the closest, then refine it with a follow-up message.
Want the AI to write the prompt for you? Try our free Photo to Prompt Generator, which turns any picture into a ready-to-use description.
Which AI Tool Should You Use?
All the major tools understand the same formula. They differ in how they like the prompt delivered.
| Tool | How to write for it | Good for |
|---|---|---|
| Google Gemini | Conversational sentences, upload the photo first | Editing uploaded photos |
| ChatGPT | Conversational sentences, follow-up fixes in the same chat | Editing and creating, easy refinement |
| Midjourney | Short descriptive phrases, add parameters such as aspect ratio at the end | Stylised and artistic images |
| Stable Diffusion | Descriptive phrases, with unwanted items in a separate negative prompt | Fine control and local use |
Features and limits change often, so check each tool's own help page for current free allowances. If you are just starting, pick one tool and learn it well before you compare others.
How to Improve a Result Without Starting Over
Beginners often throw the whole prompt away after one bad image. A better method is to treat the first result as feedback.
- Name what is wrong. Is it the light, the face, the background or the colour?
- Change only that part. Reply with "keep everything the same but make the light softer and come from the left."
- Add a missing detail. If the background looks generic, name a real place type and two objects in it.
- Lock what you like. Tell the tool "keep the pose and outfit exactly as they are."
- Save the winner. Keep a note of the prompt that worked so you can reuse it.
Our portrait prompts, men's prompts and women's prompts are good starting points for this kind of testing.
Common Beginner Mistakes
- Using vague adjectives. Words like "beautiful", "amazing" and "cool" give the AI nothing concrete. Describe light, lens and place instead.
- Stacking too many edits. Retouch, relight, change the outfit and swap the background in one request often distorts the face. Do one edit at a time.
- Forgetting the protect clause. Without it the AI may redraw your face. Always add it for edits.
- Mismatched light. A sunset background with flat indoor light on the subject looks fake. Tell the AI to relight the subject to match.
- Expecting readable text. Signs, labels and logos often warp. Keep important text out of the frame or add it afterwards in an editor.
- Copying without adapting. A prompt is a template. Replace the details with your own.
Frequently Asked Questions
What is an AI photo prompt?
An AI photo prompt is a text instruction that tells an AI image tool what to create or how to edit a photo you upload. A good prompt describes the subject, setting, lighting, camera look and mood so the result matches what you imagined.
How long should an AI photo prompt be?
Long enough to cover the five parts: subject, setting, light, camera and mood. For most beginner prompts that is between 40 and 120 words. Length itself does not help. Specific words help, and padding with repeated adjectives does not.
Do I need a paid subscription to use these prompts?
The prompts on this site are free to copy. Whether you can run them for free depends on the AI tool and its current plan, which changes over time. Check the tool's own pricing or help page for what is included.
How do I keep my own face unchanged when editing?
Add a protect clause to your prompt that says to preserve your exact facial identity, bone structure and skin tone, and not to beautify or blend the face. Also run one edit at a time. This reduces face drift, though it cannot remove it completely in every case.
Why do my AI photos look fake or plastic?
Usually the prompt asks for "flawless" or "perfect" skin, or gives no real light direction. Ask to keep pores and fine lines, give a clear light source, and add a real lens look such as 85mm f/1.8. Those three changes fix most plastic-looking results.
Can I use the same prompt in ChatGPT, Gemini and Midjourney?
Yes, with small changes. ChatGPT and Gemini accept full conversational prompts. Midjourney prefers shorter descriptive phrases with parameters at the end. Stable Diffusion works best when unwanted items go in a separate negative prompt field.
Final Thoughts
Writing a good AI photo prompt is a small set of habits, not a secret skill. Decide whether you are creating or editing. Build the prompt from five parts. Name the light and the lens. Protect your face when you edit. Then change one thing at a time until the result is right. Start with the five prompts above, keep the ones that work for you, and build your own set from there.
