This tool uses a segmentation model trained on human/portrait photos. You'll get the best results with selfie/portrait photos where the face and upper body are clearly visible — it won't give reliable results on product, animal, or general object photos.
Drag a selfie/portrait photo here or click to select
JPG, PNG, WebP — one photo at a time
Loading model…
Before
After
Try This Next
Finished here? These might be your next step.
Frequently Asked Questions
Selfie/portrait photos where the face and upper body are clearly visible. The tool uses an AI model trained for human segmentation — it can't correctly separate the background in images without people, like product photos, pets, or general objects. We don't claim it "removes the background from any photo"; the scope is deliberately limited to portrait photos.
No. Everything — including the segmentation model and the WebAssembly files — is served from openmoon.com.tr, and the processing runs entirely on your own device, inside your browser. Your photo is never sent to any server (ours or a third party's).
While processing your first photo, your browser downloads a roughly 6 MB segmentation model. This only happens on your first visit to the page (or after your browser cache is cleared) — on later uses the files come from cache, so processing starts right away.
We use the "Selfie Segmentation" model from Google's Apache-2.0 licensed MediaPipe library. The model and runtime files are hosted on our own site — no requests go out to any third-party CDN.
Last updated: