Image

Product Photography Prompt Builder

Builds image generation prompts using real photographic language — lens, lighting setup, surface, and camera position — plus a negative list and variation axes.

Prompt template

You are a still-life photographer writing a shot specification that an image
generation model will execute. You specify light and optics, not adjectives.

## Inputs
Product: {{PRODUCT}}
Material and finish: {{MATERIAL}}
Where the image will be used: {{USE_CASE}}
Brand mood in three words: {{MOOD}}
Aspect ratio: {{ASPECT_RATIO}}

## Step 1 — Shot decisions
Decide and state each of these, with one line on why it suits {{MATERIAL}}:
- Camera position and height relative to the product
- Focal length and its effect on perspective
- Aperture and the resulting depth of field
- Key light: type, size, position, and distance
- Fill and rim light, or a deliberate decision to omit them
- Surface and background: material, color, separation from the product
- Where highlights and reflections will fall

Match these to the material. Glossy surfaces need large soft sources and
controlled reflections; matte surfaces need harder light for texture;
transparent objects need backlighting and edge definition.

## Step 2 — The prompt
Write it as a single dense paragraph in this order: subject and material,
camera position, lens and aperture, lighting setup, surface and background,
color and mood, aspect ratio. Concrete nouns only.

## Step 3 — Negative prompt
List what to exclude, targeting failures specific to {{MATERIAL}} — for
example warped reflections on glass, plastic-looking skin on leather,
duplicated logo text, floating shadows disconnected from the object.

## Step 4 — Variations
Give 3 variations that change exactly one decision each from step 1, and
name the decision changed. This isolates variables so you can learn which
choice drives the result.

## Constraints
- No mood adjectives without a physical cause. Do not write "premium feel";
  write the lighting and surface that produce it.
- Never name a real brand, logo, or trademarked product design.
- Do not request legible text; generated text is unreliable. Note where text
  should be composited afterward instead.
- Do not stack more than two style references.
- State the physical setup even where the model may ignore it — the
  vocabulary shifts the output distribution regardless.

Why this works

The technique is domain-vocabulary anchoring. Image models are trained on captioned photography, so terms that appear in real photographic captions — focal length, aperture, softbox position, seamless backdrop — sit in a denser, better-conditioned region of the training distribution than vague quality words like 'professional' or 'stunning'. Step 4's one-variable-at-a-time rule turns iteration into an experiment rather than a reroll.

Example output

Example

## Shot decisions
Camera: slightly above the product, roughly 15 degrees down — shows the top
face without distorting the silhouette.
Lens: 100mm macro at f/8 — compressed perspective, product sharp front to
back, background still separated.
Key: large softbox camera-left, close, feathered — the brushed aluminum
needs a broad source to render a gradient rather than a hotspot.
Fill: white bounce card camera-right to keep the shadow side readable.
Rim: narrow strip light behind, to separate the dark edge from the backdrop.
Surface: matte charcoal seamless, product raised slightly for shadow control.

## Prompt
Brushed aluminum cylindrical desk lamp on matte charcoal seamless backdrop,
camera 15 degrees above product level, 100mm macro lens at f/8, large
feathered softbox camera-left rendering a soft vertical gradient across the
metal, white bounce fill camera-right, narrow strip rim light separating the
edge from the background, soft contact shadow beneath, cool neutral color
grade, 4:5 aspect ratio.

## Negative prompt
Blown highlights on metal, doubled or warped reflections, floating shadow,
visible text or engraving, fisheye distortion, cluttered props, HDR halos.

## Variations
1. Key light moved to hard direct source (changes: key light quality)
2. Backdrop to warm sand (changes: surface only)
3. Camera to eye level (changes: camera position only)

Usage tips

MATERIAL drives more of the output than the product itself. Glass, brushed metal, matte plastic, fabric, and skin each demand a different lighting logic, and naming the material precisely is what lets step one make sensible choices rather than generic ones.

Run the three variations before regenerating the base prompt. Rerolling the same prompt teaches you nothing; changing one named decision tells you whether the lighting or the camera position was the problem.

On text

Never ask an image model for legible packaging copy or logos. Generate the product clean and composite text in afterward — it is faster than iterating toward letterforms that will still be subtly wrong. Our roundup of AI image generators covers which models handle product surfaces best.

Browse library

Editorial Illustration Concept Brief

Generates three distinct conceptual approaches for an article illustration, each with the visual metaphor stated first and the rendering specification second.

The AI signal without the hype — new models, tools worth your time and what actually shipped. One email, no vendor pitches.

Choosing an AI stack for your team?

We help companies pick the right AI tools, wire them into existing systems and avoid the ones that quietly do not scale. Tell us what you are trying to build and we will tell you what we would use — no charge for the conversation.