Person models learn to recognize and recreate a specific individual. Once trained, you can generate that person in any style, setting, or scenario.
You only need a minimum of 1 image to train a person model. We recommend uploading several for better variety in your results — the first 3 images have the highest impact on the model's output.
Image Guidelines:
Clear face visibility in all photos
Variety of expressions and angles
Different lighting conditions
Various backgrounds (not all the same)
High resolution (at least 512x512)
Only the person you're training (no group photos)
Image order matters — if you're not satisfied with results, try reordering your training images.
Good training images:
Portrait shots (front, profile, 3/4 view)
Different expressions (smiling, neutral, serious)
Indoor and outdoor lighting
Casual and formal appearances
Full face visibility
Avoid:
Heavy makeup or costumes
Sunglasses or masks covering face
Extreme angles or blur
Photos with multiple people
Low resolution or heavily filtered images
Navigate to Models > Create New Model
Select Person as the model type
Upload your training images
Enter a name for your model
Click Start Training
Training takes around 20 seconds — you'll get a notification when it's done
Once trained, you can reuse your model unlimited times.
After training, open the dropdown below the prompt field, select your person model, and it's added to your prompt like a building block. Then describe the scene around it:
> "[your model] as a medieval knight in armor, dramatic lighting"
> "Professional headshot of [your model] in a modern office"
> "[your model] on a beach at sunset, cinematic style"
Be specific in prompts: Describe the scene, style, and mood
Match the style: Your trained person can appear in any style
Experiment: Try different scenarios to see what works best
Iterate: Refine prompts based on results
