Free Background Remover Runs Entirely in the Browser, No Photo Uploads
The author built a free background remover that processes images locally, avoiding the privacy risks of server-based tools. It is a public site forked from MIT-licensed open-source code, using the BiRefNet segmentation model via Transformers.js and ONNX Runtime in the browser. The model weights, roughly 473 MB for the default fp16 version, are downloaded directly from Hugging Face on first visit, which is the main trade-off for cold loads. Once loaded, the chosen photo is processed locally and exported as a transparent PNG without being uploaded. Because the model works at 512×512, fine edges such as hair or thin straps can look imprecise on high-resolution images.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)
Log in to join the discussion and vote.
Log in