hjLabs AI Playground
Starting the on-device runtime…
BiRefNet 512 is an ONNX model that predicts an alpha matte separating the subject from the background. On this page it runs entirely inside your browser: the 940 MB download comes straight from the Hugging Face repository “onnx-community/BiRefNet_512x512-ONNX” into your browser cache, and every inference after that executes on your own GPU through WebGPU. This is a serious download — do it once on a connection you are not paying by the megabyte for. Once it is in the Cache Storage API it stays there, and subsequent visits load it from disk in seconds.
Highest matte quality here and also the heaviest at ~940 MB — worth it only for hero images you will print. In practice people reach for a model like this for cutting out product shots and portraits into transparent PNGs, with no credits, watermark or resolution cap.
Available precisions: fp32 (full precision). fp32 comes first because at this size full precision is both the fastest WebGPU path and the best quality; the quantized variant exists for the WebAssembly fallback and for tight cache budgets. Hugging Face records about 846 downloads a month. It is a specialist choice rather than a default, which is exactly why it is worth trying when the popular model gets your particular input wrong.
Because there is no inference server behind this page, there is nothing to rate-limit and nothing to bill: the only cost of running BiRefNet 512 here is the one-off 940 MB download, and once that is cached the page keeps working with the network switched off.
birefnet 512birefnet 512 onnxbirefnet 512 browserrun birefnet 512 locallyonnx-community/birefnet_512x512-onnxremove background modelbirefnet 512 webgpubirefnet 512 download sizefree remove background modeltransformers.js model
About 940 MB for the fp32 build, including the tokenizer and config files the runtime also has to fetch — this is the whole payload, not just the weights file. It is downloaded once and then served from the browser Cache Storage API.
Yes, through the WebAssembly fallback, which is slower but works anywhere. WebGPU (Chrome or Edge 113+, Safari 26+) is what makes it feel interactive. On WASM, prefer the smaller quantized build.
No. onnx-community/BiRefNet_512x512-ONNX is a public repository, so your browser fetches the ONNX files directly over HTTPS. There is no account, no key and no server in between.
No. The model weights are downloaded from Hugging Face to your browser once, and every inference after that runs on your own GPU through WebGPU. Your text, images and audio are never sent anywhere — hjLabs.in has no inference server and no way to see your input.