← All articlesGuides for sellers

AI Photoshoot Prompt: How to Describe a Shoot That Never Happened

Megan BrooksNovember 13, 20264 min read
AI Photoshoot Prompt: How to Describe a Shoot That Never Happened

A good AI photoshoot prompt describes the entire shoot, not just a face: light, location, lens, and outfit. Here's how to structure it and what to check.

An AI photoshoot prompt doesn't describe a single image, it describes a whole set of shooting conditions. Nail those conditions down and you get a series instead of a set of disconnected images.

What the Prompt Needs to Cover

Who's in the frame. Age, type, expression, pose. No names of real people.

The look. Clothing, hair, accessories. Be specific: not "stylish," but exactly what they're wearing.

Location. Studio, street, interior, outdoors. Include details of the setting.

Light. This is the block that matters most. Soft window light, harsh sun, studio lighting with a rim light, golden hour.

Lens. A portrait lens with a blurred background, a wide angle with distortion, a long focal length with flattened perspective. Models understand these terms.

Framing. Close-up, waist-up, full body.

Building a Series Out of One Prompt

Lock down everything except pose and framing: the look, the location, the light, the lens.

Change only those two variables and you'll get a series that reads as one coherent shoot.

A reference image helps keep the person looking like themselves across generations. We cover series-consistency techniques in a separate guide.

Don't expect the face to stay perfectly identical, though. It will drift a bit from frame to frame.

What to Check in a Portrait Series

Hands and fingers: the classic spot for artifacts. There's a whole piece on why this happens.

Facial symmetry and whether the skin looks natural.

Clothing: folds, seams, fabric pattern.

Jewelry and small details.

The background behind the person. That's where the model tends to leave odd artifacts.

When the Look Isn't Coming Out Right

Describe the clothing in concrete terms. Not "business formal," but exactly what's being worn.

Cut the detail in half. An overloaded description gets simplified by the model however it sees fit.

Try a new angle before you rewrite the description. Changing the framing solves some of these problems on its own.

And check for contradictions between outfit and location. An evening gown in an artist's studio gets interpreted unpredictably.

Describing Light in One Phrase

Name the source: window, sun, studio softbox, streetlamp.

Name the direction: front, side, back, overhead.

Name the quality: soft and diffused, or hard and directional.

Three words beat a paragraph: soft side light from a window. That's usually enough to stop a frame from looking like an obvious generation.

What Should Never Be in This Kind of Prompt

Names of real people, not for the look and not as a reference for resemblance.

Imitating the style of a specific photographer. It's a gray area and tends to work worse than just describing the technique directly.

Children in any questionable context.

Requests that push against decency guidelines. Services block requests like that, and for good reason.

Any claim to documentary authenticity. A generated photoshoot is an illustration, not a real shoot, and it should never be presented as one, especially in commercial work.

I've built series like this for profile images more than once, and every time I notice I spend more words on describing the light than on everything else combined.

That's probably how it should be. Light is what separates a frame that reads as a photograph from one that reads as a generation.

FAQ

What goes into a prompt like this? Several blocks: who's in the frame, the look, location, light, lens, and framing. Light is the most important of these.

How do I get a series out of one shoot? Lock down the look, location, light, and lens, and only change pose and framing.

Will the person look identical across frames? Not fully. A reference image helps, but the face still drifts slightly.

What shouldn't go in the prompt? Real people's names, stylization after a specific photographer, and any claim that the result is a real photo shoot.

Which part of the prompt matters most? Light. It's what separates a frame that looks like a photograph from one that looks like a generation.

Sources

  1. Flami: keeping a consistent style across a series of frames
  2. Flami: why AI models mangle hands and text
  3. Flami: AI portrait from a photo

About the author

Megan Brooks

Reviewer at Flami

Read next

Get 15 credits for free

Use them to generate images and videos:

≈ 11 × Nano Banana 2, ≈ 12 × GPT Image 2, ≈ 1 × Seedance 2

Get free credits