Background Remover separates the main subject from an image and creates transparency around it. It is useful for product photos, portraits, catalogs, thumbnails, posts, composites, and files that need their background removed.
The automatic result is the starting point. Compare original and result, inspect edges, and use the mask to restore important parts or remove leftover areas.
BiRefNet Lite is the default and best option for people, products, pets, and varied objects. Smaller models are specialized: MODNet and portrait models work better for people, selfies, hair, and soft edges; fast people models are not general cutout tools for products or pets.
The tool finds what appears to be the main subject but does not fully understand photographic intent. Similar backgrounds, reflections, fine shadows, transparency, smoke, wires, multiple people, or unusual objects can need manual correction. Clearly separated subjects usually produce more direct results.
Use Protect for hair, fingers, clothing, straps, cords, accessories, and product details that must remain. Use Remove for areas that should become transparent. Use a smaller brush and high zoom around difficult edges. Changes affect only the active item's mask, not the original file.
Process the full batch or only selected images. With a large batch, start with one representative photo to confirm model and settings. Use PNG to preserve transparency. Transparent WebP also works when the destination supports it. JPEG does not retain transparency.
Start with BiRefNet Lite. It is the recommended default for people, products, pets, and varied objects.
Not necessarily. They are specialized and can be better for portraits, but were not made for general cutouts of products, pets, or complex objects.
Fine detail, colors similar to the background, transparency, reflections, and shadows can confuse the mask. Use Protect to restore the area.
A background similar to the subject can be interpreted as part of it. Use Remove to mark those regions.
Yes. Add a batch and process every image or only selected items with the same model.
PNG is the safest choice. WebP also keeps transparency when the destination app supports it.
Review hair, fine shadows, transparency, and edges at a larger size before approving the batch.
Collapse complete guide
IMAGE LAB · QUICK GUIDE
Create transparent cutouts locally and correct only the edges that actually need attention.
Always test a difficult image before applying the same setting to a large batch.
Build real-time financial projections, edit contributions directly on the chart, and see how each decision changes the future.
Edit text or images over real pages with smart placement by format and a local batch.
Split PDF batches by range, individual page, or chunks and download the results as ZIP.
Convert PDF pages to local PNG or JPEG images in batches with ZIP download.
Resize image batches by pixels, percentage, or maximum limits. Adjust aspect ratio, framing, and output format.
Calculate direct and inverse proportions for price, quantity, time, distance, and percentages.
Format, validate, and minify JSON to review API responses, configuration files, and integrations.
