hjLabs AI Playground
Starting the on-device runtime…
Toxic BERT (Detoxify) is a 109M-parameter ONNX model that classifies a piece of text by polarity or by the class it was fine-tuned on, with a confidence score. On this page it runs entirely inside your browser: the 112 MB download comes straight from the Hugging Face repository “Xenova/toxic-bert” into your browser cache, and every inference after that executes on your own GPU through WebGPU. That is a one-off download of well under a minute on a normal connection, and small enough that integrated graphics handle it without complaint. It is an English-language model; expect quality to fall off sharply on other languages.
Multi-label toxicity: insult, threat, obscenity, identity hate. A client-side comment filter that never sends text away. In practice people reach for a model like this for triaging reviews, support tickets and survey free-text without shipping customer words to a third party.
Available precisions: q8 (8-bit) and fp32 (full precision). q8 comes first because it roughly halves the download at a quality cost you will struggle to measure on this class of model; fp32 is there when you want the reference numbers. Hugging Face records about 15k 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 Toxic BERT (Detoxify) here is the one-off 112 MB download, and once that is cached the page keeps working with the network switched off.
toxic bert (detoxify)toxic bert (detoxify) onnxtoxic bert (detoxify) browserrun toxic bert (detoxify) locallyxenova/toxic-bertsentiment modeltoxic bert (detoxify) webgputoxic bert (detoxify) download sizefree sentiment modeltransformers.js model
About 112 MB for the q8 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. Xenova/toxic-bert 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.