Photorealistic Image Prompt
The difference between a generic AI image and a convincing photograph is almost entirely in the prompt. Image models were trained on vast amounts of tagged photography β if you speak photographer, they respond like one.
The Recipe
[Subject and what it's doing or showing]
Environment: [where β indoor/outdoor, specific setting]
Lighting: [type β golden hour, soft studio, harsh midday sun, neon-lit, etc.]
Camera: shot on [camera model β Sony a7R IV, Canon 5D, iPhone 15 Pro, etc.]
Lens: [focal length + aperture β 85mm f/1.4, 24mm f/2.8, 200mm telephoto]
Depth of field: [sharp throughout / shallow with bokeh background]
Style: [photojournalistic / commercial / editorial / street photography]
Mood: [describe the feeling β tense, warm, cinematic, clean and minimal]
Negative: blurry, watermark, text overlay, deformed, low quality, artificial-looking
Example prompts
Product concept:
A foldable transparent smartphone displayed on a worn wooden cafΓ© table,
morning light filtering through a window, reflections of city lights on
the glass surface. Shot on Sony a7R IV, 50mm f/1.8, shallow depth of field
with soft bokeh. Commercial product photography style, warm morning tones.
Lifestyle/portrait:
A freelance designer in her mid-30s working on a laptop at a minimal desk,
late afternoon light coming from the left, half her face in soft shadow.
Shot on Leica M11, 35mm f/2, street photography aesthetic, muted warm tones,
film grain.
Architecture/environment:
Interior of a brutalist library β raw concrete walls, tall narrow windows
flooding the space with vertical shafts of light. A lone reader at a long
wooden table. Shot on Canon 5D Mark IV, 24mm f/8, deep focus, high contrast,
editorial style.
The elements that matter most
| Element | Why it matters |
|---|---|
| Camera model | Sets the overall look β Sony vs Leica vs iPhone have distinct aesthetics |
| Focal length | 35mm = documentary, 85mm = portrait, 24mm = environmental, 200mm = compressed |
| Lighting description | More than any other element, lighting determines whether it looks real |
| Specific detail | βGold-rimmed espresso cupβ vs βcoffee cupβ β specificity reduces generic output |
| Negative prompts | Excluding watermarks, blurriness, extra fingers prevents the most common failures |
π Leftover Remixes
πΆοΈ Spicy: βGenerate three versions of this scene: one at golden hour, one at blue hour, one in harsh midday sun. Same composition, three moods.β
π§ Mild: βSame image but make it feel like a 35mm film photograph β add grain, slight color shift, natural vignette.β
π° Budget: βWhat three details should I add to this prompt to make it look less AI-generated?β
Prompt examples inspired by Amanda Caswell / Tomβs Guide. Adapted for The Prompt Kitchen.