Documentation
Generated by AI
AI PhotoMaker
Create personalized images with instant subject transformation using reference photos.
AI PhotoMaker generates new images featuring a person from your reference photos in various scenarios, styles, and poses. Upload 1-4 photos of a person, describe the desired output, and the AI creates personalized images while maintaining the subject's identity.
Core Features
- Subject-preserving image generation from 1-4 reference photos
- Multiple style presets: Cinematic, Disney, Digital Art, Photographic, and more
- Adjustable subject fidelity for identity preservation control
- "rwre" trigger word for optimal subject recognition
- Advanced parameters: steps, CFG scale, scheduler, LoRA support
- Batch generation for multiple outputs
How to Use
- Upload Reference Images: Upload 1-4 clear photos of the person you want to feature. Ensure faces are clearly visible with good lighting.
- Enter Prompt: Describe the desired scene using the "rwre" trigger word for best results. Example: "rwre person in a professional business suit, office background".
- Select Style: Choose a style preset or start with "No Style" for maximum subject fidelity.
- Adjust Subject Fidelity: Lower values (15-25) preserve more resemblance to reference; higher values (35-50) allow more creative freedom.
- Configure Advanced Settings (Optional): Adjust steps, CFG scale, and other parameters.
- Generate: Click generate and view your personalized images.
The "rwre" Trigger Word
Include "rwre" in your prompt to activate subject recognition. This trigger word tells the model to apply the reference image features to the subject in your prompt.
Example prompts:
- "rwre person as a medieval knight in armor"
- "rwre portrait in watercolor style"
- "rwre professional headshot, studio lighting"
Subject Fidelity
The subject fidelity parameter (15-50) controls identity preservation:
- 15-25: Strong resemblance to reference images, less creative variation
- 25-35: Balanced approach with moderate flexibility
- 35-50: More creative freedom, looser interpretation of reference features
Style Presets
- No Style: Raw output for maximum subject fidelity
- Cinematic: Film-like quality with dramatic lighting
- Disney: Animated character style
- Digital Art: Modern digital illustration aesthetic
- Photographic: Realistic photo-like output
- Fantasy Art: Epic fantasy-themed imagery
- Neonpunk: Cyberpunk with neon aesthetics
- And more...
Best Practices
- Use 1-4 clear reference images with good lighting
- Ensure faces are clearly visible in all reference photos
- Start with "No Style" preset for maximum subject fidelity
- Include the "rwre" trigger word in prompts for best results
- Lower Subject Fidelity values = stronger resemblance to reference
- Prompts should be 2-300 characters for optimal processing
Use Cases
- Profile Pictures: Generate professional headshots in various styles
- Creative Portraits: Place subjects in artistic or fantasy settings
- Character Creation: Design personalized character art
- Style Experimentation: See how a person looks in different artistic styles
- Social Media Content: Create unique, personalized visual content
Notes
- Reference images should have clear, well-lit faces
- Prompt length must be 2-300 characters
- Some parameters (negative prompt, CLIP skip) may not be supported by all models
- Generation may consume credits based on settings
- Download generated images for permanent storage



