Virtual Try-On

Selify AI’s virtual try-on lets your customers see how clothing looks on them before purchasing. Powered by AI image generation, it produces realistic visualizations that reduce return rates and increase purchase confidence.

How It Works

Virtual try-on uses a two-step AI pipeline:

1. Garment extraction — The AI isolates the clothing item from your product photo, understanding its shape, texture, fabric drape, and color.
2. Image synthesis — The AI composites the garment onto the customer’s photo, adjusting for body proportions, pose, and lighting conditions to produce a realistic result.

The entire process takes 15 to 40 seconds depending on image complexity and current server load. Customers receive the result directly in their Instagram DM or on your storefront.

Image Requirements

The quality of virtual try-on results depends directly on the input images.

Garment images:

• Minimum resolution: 512 x 512 pixels (1024 x 1024 recommended)
• Plain or neutral background for best results
• Full garment visible — no cropping of sleeves, hemlines, or collars
• Flat-lay or mannequin shots work best
• Avoid heavy filters or extreme color grading

Customer photos:

• Person facing forward with arms slightly away from the body
• Full torso visible (waist-up minimum for tops, full body for dresses and pants)
• Good, even lighting — avoid harsh shadows or backlit photos
• Minimum resolution: 512 x 512 pixels

Supported Clothing Types

Virtual try-on currently supports:

• Tops — T-shirts, blouses, shirts, sweaters, hoodies, jackets
• Dresses — Full-length, midi, and mini dresses
• Outerwear — Coats, blazers, vests

Items with complex layering (e.g., a scarf over a jacket) or accessories (hats, jewelry, shoes) are not yet supported. Support for pants and skirts is on the roadmap.

Integrating Virtual Try-On

Via Instagram DMs — When a customer sends a photo to your connected Instagram account with a trigger keyword (e.g., “try on”), your AI Chat Support Worker automatically:

1. Identifies the most recently discussed or tagged product
2. Runs the virtual try-on
3. Sends the result image back to the customer

Via your storefront — Use the Selify API to embed a “Try It On” button on product pages. Customers upload a photo, and results are displayed inline.

Via the dashboard — Upload garment and customer photos manually in Workforce > Model Worker > Virtual Try-On for batch processing or internal review.

Pricing

Each virtual try-on costs 20,000 micro-units ($0.20). The Free tier includes 25 try-ons per month. See How Credits Work for details on credit allocation and purchasing.

Best Practices

• Use high-resolution product photos — The higher the input quality, the more realistic the output. 1024 x 1024 pixels or above is ideal.
• Photograph garments consistently — Use the same background, lighting, and angle across your catalog for uniform results.
• Set customer expectations — Let customers know results are AI-generated visualizations, not exact representations.
• Review outputs before going live — If you’re using auto-publish, review the first 10-20 results to calibrate quality expectations.