AI virtual try-on
Choose a garment mode and an official or custom model, then set pose and background for try-on.
Virtual try-on is for garment images you want to show on a person. The workspace provides garment modes, model references, pose and background descriptions, and output sizing.
Garment modes
| Mode | Required clothing input |
|---|---|
| Matched outfit | One top and one bottom image |
| One-piece | One complete dress, jumpsuit, or similar garment image |
| Free mode | At least one garment reference |
Uploads support JPG, PNG, and WebP. Keep the garment fully visible. Enable AI matting when a complex background interferes with its outline.
Steps
- Select a mode and add the required garments.
- Choose an official model or upload a person in the custom-model tab. The custom photo becomes the active person reference.
- Enter a pose and background description; this input is required. For example: “front-facing standing pose, relaxed arms, light-grey studio background.”
- Optionally choose person features, accessories, a background template, and size.
- Sign in, check current credits, generate, then preview and download.
Compared with outfit generation
Outfit generation emphasizes quick exploration using a reference gallery. Virtual try-on offers explicit garment modes and pose/background settings for adjusting a specific product presentation.
Common questions
Why can't I proceed to generation?
Matched mode needs both a top and bottom; one-piece mode needs a complete garment; free mode needs at least one reference. You also need a model reference and a pose/background description.
How do I change the background with the same person?
Keep the model reference and change the pose/background text or select another background template before generating.