Choose an image
Upload a JPG, PNG, or WebP whose background is one solid color.
Remove an image background by picking the color to erase and tuning the tolerance, then save a transparent PNG. Runs locally on your device.
Image flows through remove background and becomes Transparent image.
Click the background color in the image to remove it.
#FFFFFF
Click another area to choose a different background color.
Three simple steps, with your content kept on your device.
Upload a JPG, PNG, or WebP whose background is one solid color.
Select the background color, then fine-tune tolerance and edge softness.
Connected background pixels become transparent and the result downloads as a PNG.
Fast, focused, and made to be clear on every screen.
Pixels close to the chosen color are removed only when connected to the image edge.
Tune the color range used to identify the connected background.
The image never leaves your device. Processing uses the Canvas API in your browser.
Open an image, click the background preview to sample a color, adjust tolerance and edge softness until the preview looks clean, and download the transparent PNG. You can also use the color picker when a precise value is easier.
Processing uses the Canvas API in the current tab, so you can remove image background online here without the picture being uploaded. That matters for product photos, ID scans, and client work that is not public yet.
The original file is untouched. Each download is a freshly encoded PNG, so nothing is overwritten and you can always start again from the source.
The tool compares pixels against the color you picked in a YCbCr color distance. It follows matching pixels from the image edges, so a similarly colored patch inside the subject is preserved. Tolerance controls how wide the match can be.
This is a color-key approach, the same idea as a green screen. It is not an AI subject-detection model, and it makes no attempt to understand what is in the picture; it only knows which pixels resemble the color you chose.
The practical consequence is that results depend on the background rather than on subject recognition. A clean, evenly lit backdrop gives the most reliable cut.
Start with tolerance low and raise it gradually. Use edge softness to blend anti-aliased pixels instead of pushing tolerance so high that the subject starts disappearing.
If a halo of background color survives around the subject, the backdrop is probably not as uniform as it looks — shadows and vignetting shift the color slightly toward the edges. Raising tolerance a little usually clears it.
Photographs of a subject against a plain wall or a paper sweep are the ideal input. JPG compression artifacts around high-contrast edges can leave speckles, so a PNG source generally cuts more cleanly than a heavily compressed JPG of the same shot.
Textured, gradient, and busy backgrounds do not have a single color to key on, so no tolerance setting will separate them cleanly. Outdoor photos with foliage or sky-to-horizon gradients fall into this category.
Hair, fur, smoke, and semi-transparent materials are still difficult. Edge softness can reduce a visible halo, but this color-based method cannot understand fine subject boundaries.
Subjects containing the background color are the third failure case. A white logo on a white backdrop will lose the logo along with the background, because the tool has no way to tell them apart.
Product photography is the most common use. A transparent cutout drops onto any listing page, catalog, or slide background without a visible box around it, which is why marketplaces so often require it.
Logos and signatures are the other major case. Using this as a transparent background maker turns a scanned signature or a logo exported on white into something you can place over colored headers and documents.
PNG is the format that carries transparency here, so to make PNG background transparent is exactly what the download gives you. Note that converting that PNG to JPG later discards the transparency and fills it with a solid color, since JPG has no alpha channel.
The Image Cropper is useful before removal, because trimming away irrelevant edges reduces the amount of background you need to key out in the first place.
The Image Color Extractor helps identify the exact background shade when it is not obvious by eye, and the Image Metadata Viewer shows what a file records about the camera and capture settings.
All of these tools process images locally in the browser, so a full crop-inspect-remove pass keeps every version of the picture on your own device.
It compares colors in a YCbCr distance, follows matching pixels from the image edges, and makes that connected region transparent. It works best on clean, solid backgrounds.
No. This is a simple local color-key tool — there is no AI model and no network request. Complex subjects or textured backgrounds may need manual editing.
A PNG with transparency. Images up to the practical output limit keep their original dimensions; very large images are processed at a smaller working size to protect browser memory.