Prompts / Multimodal

Visual Multimodal & Vision Prompts

Master the art of image generation and analysis. From cinematic DALL-E 3 descriptions to precise photo editing instructions for AI agents.

Cinematic Photorealism Generator

Use Case: Generating high-fidelity, cinematic image prompts for DALL-E 3 or Midjourney.
Prompt
### ROLE
Act as a world-class Cinematographer (Director of Photography).

### TASK
Convert the simple idea "[INSERT IDEA, e.g., a cyberntic cat]" into a professional image generation prompt.

### PARAMETERS TO DEFINE
1. **Subject**: Detailed physical description, texture, and pose.
2. **Lighting**: Type (e.g., volumetric, rim, chiaroscuro), color temperature, and source direction.
3. **Camera Gear**: Lens focal length (e.g., 85mm for portraits), aperture (f/1.8 for bokeh), and film stock aesthetics (e.g., Kodak Portra 400 grain).
4. **Composition**: Rule of thirds, leading lines, or symmetry.
5. **Style**: "Hyperrealistic, 8k resolution, Unreal Engine 5 render style".

### OUTPUT
A single, dense paragraph optimized for DALL-E 3 digestion.

Why it works with GPT-5.2

Targets 'chatgpt prompt for cinematic photo'. By specifying camera gear and lighting terminology, it forces the image generator to adopt a specific aesthetic, avoiding the 'AI plastic' look.

Expected Output

A rich, descriptive paragraph like: 'Close-up portrait of a cybernetic cat, 85mm lens, f/1.8, soft morning light streaming through a window, visible dust motes, intricate metallic fur texture...'

Advanced Variation

Change Style to 'Studio Ghibli style, watercolor texture, vibrant cel shading' for the anime intent.

The AI Photo Editor Agent

Use Case: Instructing a multimodal AI to edit an uploaded image specifically.
Prompt
### ROLE
Act as a Senior Photo Retoucher using Photoshop.

### INPUT
I have uploaded an image.

### TASK
Perform the following edits:
[INSERT EDIT REQUEST, e.g., "Remove the background and color correct for a warm sunset vibe"].

### CONSTRAINTS
- **Precision**: Do not alter the subject's facial features or identity.
- **Lighting Consistency**: If adding a new background, ensure the shadow direction on the subject matches the new light source.
- **Output**: Return the edited image and a text summary of the changes made.

Why it works with GPT-5.2

Targets 'chatgpt prompt for photo editing'. It sets clear boundaries (Constraints) to prevent the AI from hallucinating unwanted changes to the subject, which is a common issue in AI editing.

Expected Output

An edited image file where the background is removed/replaced, plus a log saying 'Matched subject lighting to new sunset background (warm orange tint applied)'.

Advanced Variation

Task: 'Analyze the composition of this photo and suggest 3 cropping variations to improve visual impact'.