How to prompt GPT Image 2 for product shots
Studio-quality product photography from a prompt — surface, lens, lighting, and the one detail most prompts skip.
The structure
Every product-shot prompt that works has five parts:
- Surface — what the product sits on, what's around it
- Product — generic description, materials, finish
- Lens — focal length, aperture, framing
- Lighting — direction, quality, color temperature
- Grounding detail — the one imperfection that sells it as real
Working template
Studio product photograph, 1:1 square. A matte black ceramic coffee mug sitting on a raw concrete surface, a single dried eucalyptus sprig casting a soft shadow nearby. 85mm macro lens, f/2.8, mug centered, shallow depth of field — surface texture sharp, background falls off softly. Soft directional window light from camera-left, warm afternoon temperature, gentle fill from a white bounce on the right. Faint wisp of steam rising from the mug, near-imperceptible water ring on the concrete. Editorial, Kinfolk-magazine quality. No text.
Why each part matters
Surface
The surface does half the work. Concrete, marble, raw oak, weathered linen, brushed steel — name it specifically. "A nice background" gives you stock-photo mush.
Add one or two contextual props (eucalyptus sprig, folded linen, a small ceramic dish). They make the frame feel intentional.
Product
Be generic but specific. Materials and finish over brand. "Matte black ceramic," "brushed aluminum," "natural oak with visible grain." The model can't render real logos — don't fight it.
Lens
This is where amateur prompts fall apart. Use real photographic vocabulary:
| Use case | Prompt |
|---|---|
| Hero packshot | "85mm lens, f/2.8, centered" |
| Lifestyle scene | "35mm lens, f/4, slight angle" |
| Macro detail | "100mm macro lens, f/2.8, extreme close-up on [detail]" |
| Flat lay | "50mm lens, f/8, top-down" |
Lighting
Name the direction and the quality:
- "Soft directional window light from camera-left"
- "Hard noon sun, hard shadows"
- "Diffuse softbox above, gentle fill below"
"Good lighting" does nothing. The model has no opinion on good — it has opinions on directional.
Grounding detail
The thing every AI product shot is missing. Add one:
- Steam rising from a cup
- Condensation on a bottle
- A faint shadow under a flat-lay object
- A subtle fingerprint smudge on glass
- A soft reflection on polished metal
One imperfection. That's what separates rendered from photographed.
Skip the rewrite
Depikt's generator detects product-shot intent and applies this structure automatically, including the grounding-detail line that most prompts skip.
Generate yours
Generate polished prompts in seconds.
Paste a rough idea. Get back a structured prompt that ships.
More in How-to
How to prompt GPT Image 2 for logos and brand marks
GPT Image 2 can spit out genuinely usable logo directions if you brief it like a designer. Here's the four-block structure: concept, mark type, type system, finish — plus copy-paste templates.
How-toHow to prompt GPT Image 2 for infographics and diagrams
GPT Image 2 can render genuinely usable infographics — but only if you treat the prompt like a brief, not a vibe. Here's the five-block structure: format, hierarchy, data, type system, and finish.