JSON Prompt Builder

Model-aware output · No API · In-browser

Describe a shot in plain English — it auto-detects subject, style, lighting and optics, then formats the prompt for your target model: Midjourney V8.2 flags, FLUX 3 Image positive prose, separate SDXL prose + tags + negative, or structured JSON for Nano Banana Pro / Gemini. Runs locally, no key required.

Image Description No subject detected
Styles
Film Stocks
Optics
🎨 Scene & Style
📷 Camera & Optics
💡 Lighting
⊞ Composition
◈ Materials & Surface
✦ Quality & Rendering
Rendering Options
◎ Character Consistency
Keeps a character identical across multiple generations. Leave blank to omit from JSON.
JSON Output

Model-specific prose uses the shared CinePrompt profiles. Generic and Gemini JSON are creative briefs, not API request payloads. Select output size and attach references in the destination tool.

{ } Start typing above to generate your JSON prompt Auto-detects subject, style, lighting, weather, and more from your description.
Tags will appear here once a prompt is detected.