RMBG is a cross-platform image background removal application that uses multiple open-source AI models to separate foreground from background. It runs entirely on the local device, avoiding server uploads, and supports batch processing of up to twenty images with a maximum size of five megabytes each. Users can select from various matting models for different image categories.
RMBG was created as an open-source project focused on private, local background removal. The application has been developed since the early 2020s, when open-source matting models such as U2-Net and ISNet became available for public use. The developers are a community of contributors rather than a single commercial company. They built RMBG using web technologies, including JavaScript, HTML, and CSS, with AI inference performed through browser-compatible runtimes.
