Design the shot before AI generates it
Turn your subject, UI reference, and scene idea into a structured visual prompt for Claude, Gemini, and ChatGPT. More control, less guesswork, fewer random results.
Preserve identity, pose, and realism
AI is powerful — but unpredictable
One-line prompts often distort faces, ignore layout references, change outfits, or break composition. AI Scene Director helps you structure the visual before generation.
Direct the scene in layers
This tool is built to make prompting feel like creative direction, not guesswork.
Lock the subject
Upload a person and preserve identity, pose, and realism in a more controlled way.
Map the screen
Use any UI or mockup as the exact phone display reference for your prompt.
Control the environment
Choose location, lighting, camera style, wardrobe direction, and scene tone.
Generate model-ready prompts
Get structured outputs adapted for Claude, Gemini, ChatGPT, and a full power version.
Built for creators, mockups, and ad visuals
Use it for polished visual directions instead of writing random prompts from scratch every time.
Influencer mockups
Create oversized phone scenes, app visuals, and editorial social mockups.
Fitness promotions
Build premium gym, wellness, athleisure, and movement campaign visuals.
Skincare ads
Structure clean beauty visuals with product mood, wardrobe, and scene control.
Streetwear campaigns
Keep the visual tone consistent while exploring city scenes and fashion styling.
App launch visuals
Showcase a UI naturally inside a scene without losing perspective or composition.
Creative direction boards
Use the tool as a structured shot planner before testing in multiple AI models.
From rough idea to structured AI shot
Simple enough to use quickly, detailed enough to produce stronger scene instructions.
1. Upload references
Add a subject image and a screen or UI reference to define the core visual anchors.
2. Direct the scene
Choose pose, city, styling, lighting, and camera behavior like a creative director.
3. Generate prompt
Copy the optimized output into Claude, Gemini, or ChatGPT and start testing immediately.
Builder preview
This HTML version is a polished front-end concept for your tool. It is designed to feel premium even without API-based generation.
Start with a preset
Lock the person and pose
Map the screen reference
Control environment, lighting, and tone
Separate the look from the subject reference
Define how the shot should feel
Generated Scene Direction
Build one creative direction, then copy the version best suited for your preferred AI model.
Generate a highly realistic editorial photograph in Dubai where the subject from the uploaded identity reference stands beside an oversized vertical smartphone displaying the uploaded UI reference exactly on screen. Preserve identical facial structure, proportions, skin texture, and hairstyle from the subject reference. Do not beautify, stylize, or alter age. The scene should feel like a luxury editorial campaign shot on an authentic Dubai street with the Burj Khalifa faintly visible in the distant skyline. The subject is standing upright beside the oversized phone with a neutral confident expression, front-facing body orientation, and one hand resting naturally on the top edge of the phone. Wardrobe should be premium athleisure in a cream, charcoal, and muted beige palette, clearly different from the clothing in the identity reference. Use sunset lighting, eye-level frontal composition, realistic grounded shadows, 35mm editorial lens behavior, and moderate background blur. Keep the phone aligned to street perspective with natural screen reflections, subtle glare, metallic frame realism, and a strong grounded shadow on pavement. Do not cartoonize, stylize, or surrealize the scene. Prioritize photorealism, high resolution, visual consistency, clean composition, and accurate UI mapping.
Stop guessing your prompts
Use AI Scene Director to structure the shot first, then test the output inside Claude, Gemini, or ChatGPT with much stronger control.