remove people from photos

How this people-removal tool works

This tool does not detect a person. You choose a local image, paint a mask over the person or the parts to remove, and run the existing browser-side reconstruction.

The reconstruction fills masked pixels by looking for similar unmasked patches elsewhere in the same image and copying their color values into the selected area. It is a patch-based pixel repair, not a generative model, semantic person remover, or recovery of the original hidden scene.

Keep the mask close to the subject, but include stray edges that should disappear. The tool slightly expands the painted mask to catch edge pixels, so a selection that touches a nearby feature can alter those adjacent pixels too. You can refine the mask and rerun from the original image before downloading.

“Remove person from photo,” “erase people from photos,” “delete people from photo,” “erase person from photo,” and “delete person from photo” describe the same manual edit here. The phrase “remove person from photo online free” also refers to this browser-based tool, not a separate processing mode. Other ways to describe the same request include “erase someone from a photo,” “remove people from picture,” and “remove people from photos free.” None of these wordings switches on person recognition or changes how pixels are reconstructed.

Which photos are suitable for a manual repair?

• Flat or repeating background: usually the most workable case because nearby unmasked pixels can provide a closer texture match.
• Small person or isolated part: a smaller mask leaves more surrounding image available as a reference, but the result still needs visual inspection.
• Overlapping people: difficult when one person covers another; the hidden boundary and detail are not present in the file for the tool to restore.
• Hair and fine edges: easy to under-select or include nearby background. Paint carefully and check the enlarged comparison for halos or broken strands.
• Large person, large mask, or complex texture: likely to show mismatched patches, repeated patterns, or altered nearby details; this page cannot promise a seamless result.

Before starting, look at what remains around the person. If the selected area covers most of the frame or the background has unique detail, the tool may not have a close match to borrow. Keep the original file and compare the full image as well as the enlarged crop.

File handling and boundaries

• Choose one JPG/JPEG, PNG, or WebP image up to 20 MB. The edited image is downloaded as PNG.
• The selected image is read and reconstructed in this browser; the image-processing code does not send it to a processing server. Separate analytics or advertising scripts may load.
• You paint the region yourself. There is no automatic person detection, multi-file batch, or hidden-scene recovery.
• Large image dimensions can use substantial device memory. The file-size limit does not guarantee that every device can process every image.

Use this workflow only on photos you own or are authorized to edit. If you need the real scene behind the person, an unflattened original or a new photograph is required; this pixel repair cannot reveal content that the image does not contain.

Frequently asked questions