useknockout/api

useknockout is an open-source background removal API, a service you call with an image file and get back the same image with the background stripped out, leaving a transparent PNG. The problem it solves is a common one in product photography, app design, and content creation: cleanly removing backgrounds from photos, especially from tricky subjects like hair, fur, or complex edges. Under the hood it uses an AI model called BiRefNet, which the README describes as state-of-the-art on standard benchmarks for object segmentation, meaning separating a subject from its background. The service is hosted on GPU infrastructure and processes each image in roughly 200 milliseconds on a warm server. You call it over HTTP by sending an image file and receiving a processed image back. Beyond basic background removal, it offers several ready-made presets: you can replace the background with a solid color or another image, create sticker-style images with a white outline, generate e-commerce studio shots with a white background and drop shadow, tightly crop to the subject, or produce a side-by-side before-and-after image. Batch processing up to 10 images per call is also supported. You would use this when building a product that needs image cutout capabilities, such as an e-commerce site, a photo editing tool, or any app where users upload photos, and you want to self-host the solution rather than paying per-image fees to commercial services. The backend is written in Python with a TypeScript SDK available. It is MIT licensed and currently free during its public beta.