Briefs, not magic words
AI photoshoot prompts you can adapt
PublicSelf editorial · Published
A useful AI photoshoot prompt describes a photograph, not a mood. Name the person reference, the outfit, the pose, the place, the light and the camera view, and keep those choices compatible: a close-up cannot show a whole outfit, and a busy street can swallow a small profile portrait.
Treat the five prompts below as starting briefs. In PublicSelf you can paste or adapt them in Describe mode, which takes a description of the complete shot. The text does not attach anything by itself: your identity and your Look still supply the person and the clothes, so choose those first. You can also skip writing entirely and set the same choices with the Shoot controls.
1. A clear professional portrait
Prompt: “Waist-up portrait. Relaxed upright posture, shoulders slightly turned, looking at the camera with a gentle closed-mouth smile. Plain light-grey studio background, soft window light from the left, camera at eye level. Keep the face and collar clearly visible.”
Use this when being recognised matters more than a striking setting. A knit, an open-collar shirt or a tailored jacket can all work; pick what suits your work and audience. Square (1:1) suits most profile slots. Check the face, collar edges and shoulders, then test the image at thumbnail size with room around the hair.
Next: compare the professional headshot page with personal branding photos, which covers a wider set of pictures.
2. A full-length street-style photograph
Prompt: “Full-length street-style photograph. Long charcoal coat over a cream knit, straight dark trousers and black loafers. Relaxed standing pose beside a stone shopfront, overcast afternoon light, camera at eye level. Show both shoes completely with a little pavement beneath them. Keep signs and passers-by unobtrusive.”
The outfit and the street should both contribute. Build the coat, knit, trousers and loafers into your Look first, because a description alone does not select garments. Start standing so the silhouette is easy to inspect; a walking version adds harder foot placement and fabric movement. Portrait (9:16) gives a full-length figure room.
Next: street-style photoshoot ideas cover other settings and eras.
3. Restyle one part of a Look
Prompt: “Same person and outfit, but with a dark olive overshirt instead of the jacket. Cream top, dark trousers and simple black shoes stay as they are. Full-length against a quiet background so the layers and proportions are easy to compare.”
Make the actual change in the Look builder: adding a piece to a slot replaces what was there, which is more reliable than asking in words. Use the prompt to keep the rest of the frame calm. Review the whole result, because other garment details can still shift during generation.
Next: remix outfits with AI for the wardrobe workflow and outfit color combinations for palette choices.
4. A portrait with one accessory
Prompt: “Three-quarter portrait. A simple dark tote bag hangs from one relaxed hand beside the body. Quiet café exterior, soft daylight, camera at eye level. Leave space around the hand, handle and bag so their connection is visible.”
Add the bag under Props & items, from the catalogue or by uploading a product photo, and describe how it is held; the prompt alone does not create the object reliably. Inspect fingers, handle, wrist and the bag's size. If they merge or the grip looks impossible, simplify the pose or drop the object. A product photo helps describe an item but does not guarantee exact branding or construction.
Next: read adding props to AI photos before trying more complicated interactions.
5. An editorial portrait with room for a headline
Prompt: “Landscape editorial portrait. Person in the left third of the frame, wearing a sculptural dark jacket over a simple top. Waist-up, calm expression, textured pale wall and soft directional light. Keep the right side quiet for text added later.”
Choose Landscape (16:9) before you generate. Keep the empty area deliberate and check that the face stays large enough for the final layout. Add headlines in your layout tool afterwards. On a phone, a narrow screen may need the whole image above the text rather than a hard crop.
Next: editorial photoshoots and portrait camera angles.
Improve a prompt after looking at the result
Name the visible problem. “Show both shoes with space below them” fixes a cropped outfit. “Keep the bag beside the leg with the handle fully visible” fixes an unclear prop. Adding more style adjectives rarely fixes either.
Keep the brief next to the image so you can compare what you asked for with what appeared. Reuse the parts that worked and change one decision at a time. Review identity, clothing and anatomy on every result, including variations made from the same references.
- Name the framing: close-up, waist-up, three-quarter or full-length.
- Name one light source and its direction.
- Describe the place in a few concrete nouns.
- Say what must stay visible: shoes, hands, a held object.
- Choose the format before generating.
Before you generate from a prompt
- Identity and Look are selected; the text does not replace them.
- The framing can actually show what the prompt describes.
- Props are added under Props & items, not only named in the text.
- The format matches where the picture will be used.
- You know the price on the Generate button.
Questions about AI photoshoot prompts
Do these prompts guarantee the exact picture?
No. They describe a target. Your references, the selected Look and controls, and the model all affect the result, and fine details can vary from one generation to the next.
Where do I paste a prompt in PublicSelf?
Use Describe mode in Shoot, which takes a description of the complete shot: scene, pose, expression, lighting, framing and mood. On a phone it appears as Write your own shot. Choose your identity and Look first, because the text does not select garments or attach references.
Can I skip writing a prompt?
Yes. Pick a Look, then set pose, expression, framing, scene and light with the Shoot controls. The briefs on this page can still help you decide what those details should be.
Why does my prompt ignore part of the outfit?
The Look decides the clothes. If a garment matters, put it in the Look rather than only naming it in the prompt, and make sure the framing is wide enough to show it.
Can I publish the generated result?
Review the image for its intended use, use references you are allowed to use, and follow the destination's rules. A generated scene should not be presented as evidence that an event really happened.