How customer photos are handled
We take customer privacy seriously. Here's what happens when a customer uploads a photo for virtual try-on.
Photo processing flow
Upload: Customer uploads their photo via encrypted HTTPS connection
Consent check: If you've enabled the consent requirement, customers must agree before upload
Processing: The photo is sent to the AI model along with the garment image
Result: The generated try-on image is returned and stored for the customer's history
Retention: Photos are kept based on your settings (auto-delete after 12 months if enabled)
Content moderation
MirrorNex includes content moderation to filter inappropriate uploads. You can configure the moderation level in Widget Settings → Behaviour:
- Conservative — Blocks swimwear, underwear, and revealing outfits. Use for children's clothing or culturally conservative stores.
- Standard — Allows swimwear and underwear, blocks explicit content. Recommended for most fashion stores.
- Off — No content filtering. Only for swimwear, lingerie, or activewear stores.
What we DON'T do with photos
- ❌ We do NOT use photos to train AI models
- ❌ We do NOT sell or share photos with third parties
- ❌ We do NOT use photos for any purpose other than generating try-on results
Data security
- ✅ All uploads are encrypted with TLS 1.3
- ✅ Photos are stored in encrypted cloud storage (R2)
- ✅ Automatic deletion based on your retention settings
- ✅ Customers can delete their data via the Customer Portal
Customer control
Customers have control over their data through the Customer Portal:
- View history — See all their past try-ons
- Delete photos — Remove individual try-ons or all their data
- Export data — Download all their data as a JSON file (if you've enabled this)
- Manage consent — Update their privacy preferences
BYOK (Bring Your Own Key) users
If you use your own API key:
- Photos are sent to your chosen AI provider (Fashn.ai)
- The provider's data handling policies also apply
- MirrorNex still stores photos for customer history based on your settings