The 5-part prompt formula is the fastest way to stop getting random AI photos. You write five short pieces in a fixed order: Subject, Style, Lighting, Camera and Mood. This guide explains each part with ready-to-use words, a worked example, and four free copy-paste prompts for ChatGPT, Gemini, Midjourney and Stable Diffusion.
Table of Contents
Key Takeaways
- Write every prompt in the same order: Subject, Style, Lighting, Camera, Mood.
- Every part you skip is a part the AI guesses, and guesses cause random results.
- Lighting and camera words change a photo more than extra adjectives do.
- Change one part at a time and re-run, so you learn which part did the work.
Why a Formula Works
An AI image tool fills every gap in your prompt with its own guess. Write "a woman in a park" and it picks her age, her clothes, the time of day, the lens and the colour of the whole image. Sometimes the guess is good. Most of the time it is average, because the AI aims for the most typical answer.
A formula removes the gaps in a fixed order. You decide the five things that matter most, so the AI guesses only the small details. It also makes your work repeatable. When a prompt works, you know exactly which five pieces made it work, and you can reuse them on a new subject.
This page goes deeper than the complete beginner guide to AI photo prompts. If you are new to prompts, read that guide first, then come back here.
The Formula at a Glance
| Part | Question it answers | Example words |
|---|---|---|
| 1. Subject | Who or what, and doing what? | a fisherman in his fifties, wool cap, mending a net |
| 2. Style | What kind of image, and where? | realistic documentary photo, lakeside dock, misty mountains behind |
| 3. Lighting | Where does the light come from? | soft sunrise light from camera left, gentle rim light |
| 4. Camera | Lens, distance, angle, focus | 50mm lens, f/2, eye-level, waist-up shot |
| 5. Mood | Feeling, colour and finish | calm, muted earth tones, natural skin texture, sharp focus |
Note: In our complete beginner guide the second part is called "setting". Here we call it "style" and include the setting inside it, because the look of the image and the place it happens in are usually written together.
Part 1: Subject
The subject is the person, animal or object, and what they are doing. Start here because most tools give the strongest weight to the first words of a prompt.
A strong subject has four details: who (age range, gender if it matters), what they wear, what they do, and where they look. Compare these two:
- Weak: "a man"
- Strong: "a man in his forties wearing a charcoal wool coat, laughing, looking slightly off camera"
The second version gives the AI an expression and a gaze direction, which are the two details that make a face look natural instead of posed.
Tip: Describe one subject well before you add a second. Two people in one prompt double the chance of merged faces or wrong hands.
Part 2: Style
Style tells the AI what kind of image to make and where the scene takes place. Name the image type first, then the place.
| Style word | What you get | Best used for |
|---|---|---|
| realistic photograph | A natural camera look with normal colour | Portraits, profile photos, lifestyle |
| studio portrait | Plain backdrop, controlled light | Headshots, product shots |
| cinematic film still | Movie colour grade, wide framing | Dramatic scenes, storytelling |
| documentary street photo | Candid, slightly imperfect, real places | Travel and everyday life |
| editorial fashion | Clean, styled, high contrast | Fashion and magazine looks |
After the image type, add the place and the background: "on a wooden lake dock, misty mountains blurred behind". The words "blurred behind" matter. They tell the AI to keep the background soft so the subject stays the focus.
Part 3: Lighting
Lighting is the part beginners skip, and it is the part that makes a picture look professional. Always name two things: the direction of the light and its quality.
| Words to use | Effect |
|---|---|
| golden hour sunlight from camera right | Warm, directional light with long soft shadows |
| soft window light from the left | Gentle, flattering indoor portrait light |
| overcast daylight | Even light with almost no hard shadows |
| rim light behind the subject | A bright edge that separates the subject from the background |
| neon light, cyan and magenta | Night-city colour with coloured reflections |
"Nice lighting" is not a lighting instruction. The AI cannot place a light from that phrase. "Soft window light from the left" tells it exactly where the brightest side of the face should be.
Part 4: Camera
The camera part sets the lens, the distance, the angle and the depth of field. You do not need to be a photographer. Four details are enough.
- Lens: 35mm is wide and environmental, 50mm looks natural, 85mm is the classic portrait lens.
- Aperture: f/1.8 gives a very blurred background, f/8 keeps more of the scene sharp.
- Angle: eye-level feels honest, low angle feels powerful, high angle feels small or gentle.
- Crop: close-up, head-and-shoulders, waist-up or full-length.
These words describe a look. The AI is not using a real lens, so the result will not match a real 85mm photo exactly. It still moves the image in the right direction, which is why the words work.
Part 5: Mood
Mood is the feeling of the picture and its finishing quality. This is where you add colour, emotion and realism details.
- Feeling: calm, joyful, serious, mysterious, nostalgic.
- Colour: warm tones, muted earth colours, teal and orange, soft pastel.
- Finish: natural skin texture, visible pores, sharp focus on the eyes, fine film grain.
For realistic people, always ask for "natural skin texture" or "visible pores". Without it, many tools smooth the skin until it looks like plastic.
Build a Prompt in 5 Steps
Use this exercise to see the formula work. Start with a weak prompt and add one part at a time.
- Subject: "A fisherman in his fifties wearing a wool cap, mending a net, looking down."
- Add Style: "...realistic documentary photograph on a wooden lake dock, misty mountains blurred behind."
- Add Lighting: "...soft sunrise light from camera left with a gentle rim light."
- Add Camera: "...50mm lens, f/2, eye-level, waist-up shot."
- Add Mood: "...calm, muted earth tones, natural skin texture, sharp focus on the hands."
Run the prompt after each step and save the results. The jump between step 1 and step 5 is the clearest proof of what each part does, and it is the best test you can show your readers.
4 Copy-Paste Prompts
Each prompt below uses all five parts. The square brackets are only labels, so delete them if your tool does not like them. Replace the example details with your own.
1. The Fisherman Portrait (Create)
The five-step exercise above, assembled into one prompt.
Prompt 01
Best for: ChatGPT · Gemini · Midjourney
A fisherman in his fifties wearing a wool cap, mending a net, looking down [subject], realistic documentary photograph on a wooden lake dock with misty mountains blurred behind him [style], soft sunrise light from camera left with a gentle rim light [lighting], 50mm lens, f/2, eye-level, waist-up shot [camera], calm mood, muted earth tones, natural skin texture, sharp focus on the hands [mood]. Ultra-realistic. 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. Studio Headshot (Create)
A clean professional look for a profile picture.
Prompt 02
Best for: ChatGPT · Gemini · Stable Diffusion
A woman in her thirties wearing a dark blazer and white shirt, confident relaxed smile, looking at the camera [subject], studio headshot against a plain light-grey backdrop [style], large soft key light from camera left with a subtle fill on the right [lighting], 85mm lens, f/2.8, eye-level, head-and-shoulders crop [camera], professional and approachable, neutral colours, natural skin texture, sharp focus on the eyes [mood]. Ultra-realistic. Free from distorted features, 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. Cinematic Night Street (Create)
Strong colour and contrast, good for testing how lighting changes the mood.
Prompt 03
Best for: Midjourney · Gemini · ChatGPT
A young man in a black rain jacket standing at a crosswalk, hands in pockets, looking toward the camera [subject], cinematic film still on a rainy city street at night with wet road reflections [style], cyan and magenta neon light from signs on both sides with a soft rim light [lighting], 35mm lens, f/1.4, low angle, full-length shot [camera], moody and quiet, teal and magenta tones, fine film grain, sharp focus on the face [mood]. Free from distorted hands, warped text and extra limbs. 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. Edit Your Own Photo (Edit)
Upload your own photo first. The formula here describes only what changes, and the last sentence protects your face.
Prompt 04
Best for: ChatGPT · Gemini (upload your photo)
Edit the uploaded photo. Keep my exact face, bone structure, skin tone and expression, and do not beautify or blend the face [subject]. Keep my outfit and pose unchanged, and change the background to a quiet park path with soft green blur [style]. Relight the scene with warm golden-hour sunlight from camera right and a soft rim light on my shoulder [lighting], 85mm portrait look, shallow depth of field, same crop as the original [camera], warm natural tones, keep real skin texture and pores, sharp focus on the eyes [mood]. Do not alter my facial identity in any way. 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.
When Two Parts Fight Each Other
A strong prompt has five parts that agree with each other. Beginners often write parts that fight, and the AI then averages them into something muddy. Watch for these pairs:
- Soft overcast light + deep black shadows. Overcast light has almost no hard shadows.
- Cheerful mood + night rain city. The scene says dark and wet, the mood says bright. Pick one.
- Wide 24mm lens + extreme background blur. Wide lenses keep more of the scene sharp.
- Studio backdrop + outdoor sunlight. Choose the place first, then the light that fits it.
Common Mistakes
- Writing the parts in a random order. A fixed order makes prompts easy to edit and compare.
- Stacking adjectives. "Beautiful, stunning, amazing, perfect" adds nothing. One concrete lighting word is worth more.
- Changing everything between runs. You learn nothing. Change one part, then run again.
- Forgetting the protect clause when editing. The AI may reshape your face without telling you.
- Copying a prompt without adapting it. A prompt is a template. Replace the details with your own.
Frequently Asked Questions
What is the 5-part prompt formula for AI photos?
It is a simple order for writing a photo prompt: Subject, Style, Lighting, Camera and Mood. Each part answers one question about the picture, so the AI has fewer things to guess.
Does the order of the five parts matter?
A little. Put the subject first because tools give the most weight to the start of a prompt. The other four parts can move around, but keeping one fixed order makes your prompts easy to edit and compare.
Do I need all five parts every time?
No. For quick tests, subject and lighting are enough. Use all five when you want a result you can repeat or publish. If you leave a part out, the AI chooses it for you, and that is where random results come from.
What is the difference between style and mood?
Style is the look of the image: realistic photo, film still, illustration, studio shot, and where the scene takes place. Mood is the feeling and finish: calm, dramatic, warm tones, soft contrast, natural skin texture.
Can I use this formula to edit my own photo?
Yes. Keep the same five parts, but describe only what should change, and add a sentence that protects your face and identity. Our beginner guide explains the protect clause.
Why does my result still look wrong when I use all five parts?
Usually one part is too vague, or two parts disagree, such as soft overcast light with hard black shadows. Change one part at a time and run it again, so you can see which part caused the problem.
Final Thoughts
The 5-part formula is a checklist, not a rule book. Write the subject, name the style, place the light, set the camera and finish with the mood. When a result disappoints you, ask which of the five parts was vague, fix that part, and run it again. Use the four prompts above as a starting set, and build your own library as you learn what works for your tool.
