Bring your photo.
A portrait, a product, your favourite plant. Choose a JPG, PNG or WebP.
Turn a photo into a possibility. Remove the background,
find your perfect canvas, and make something yours.
Keep the good stuff.
AI runs on your device. First use downloads a model (about 98 MB Fast or 473 MB Detailed); your browser may cache it for later. Your photo stays on this device.
or drop an image anywhere here
Paste a photo with Ctrl + V
Preparing your photo…
First use downloads an AI model (about 98–473 MB).
Processing speed depends on your device.
A portrait, a product, your favourite plant. Choose a JPG, PNG or WebP.
AI separates the subject on your device. Check the result and refine the edges.
Keep it transparent or add a fresh background. Download a PNG, ready to use.
The AI runs on your device after your browser downloads the model. That avoids a per-image AI server fee, while your photo stays on your device.
₹0 / month
₹99 / month
Browser-only limits can be bypassed. Strict caps require server-side processing, which changes the privacy and operating-cost model.
Your next great image shouldn’t need an account—or a trip to someone else’s server.
Clearcut runs the AI in your browser. Your chosen photos and edits stay in this tab. Model files download from external hosts, which receive normal network request information, but your photo is never attached.
Find your fresh start ↗This version is free, with no watermark or account requirement. Google sign-in is optional and does not change processing or unlock quotas today. Processing uses your device, so there is no per-photo AI API charge. Hosting and bandwidth still have costs and limits.
No image upload service, analytics, or image storage is used. Your photos, masks, and edits remain in browser memory. Clearing the photo or closing this tab discards them. AI models may be cached by your browser for future visits.
The first run downloads a model of approximately 98 MB, or up to 473 MB in detailed mode. Browser caching can help subsequent runs. Older devices may be slow or run out of memory. Use a current Chrome or Edge browser for the initial release.
No. Hair, glass, shadows, low contrast, and multiple subjects can need manual refinement. Both models work at 512px internally; the output can keep the photo’s dimensions, but that does not add native-resolution AI detail. This release targets desktop Chrome and Edge; iOS is not supported yet.
The engine and configured models use Apache-2.0 and MIT licenses. You are responsible for the rights to the photos you choose. See the included third-party notices for component attribution.