hjLabs AI Playground
Starting the on-device runtime…
CLIP ViT-L/14 is a 428M-parameter ONNX model that scores an image against arbitrary text labels by comparing image and text embeddings. On this page it runs entirely inside your browser: the 436 MB download comes straight from the Hugging Face repository “Xenova/clip-vit-large-patch14” into your browser cache, and every inference after that executes on your own GPU through WebGPU. Expect a download of roughly a minute on a decent connection. It fits comfortably in the browser cache and in the memory of any GPU from the last few years, including laptop integrated graphics.
The strongest CLIP here and the one used by most diffusion pipelines; 436 MB and noticeably slower. In practice people reach for a model like this for searching a photo library in plain language and tagging images against a taxonomy you invent on the spot.
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 17k 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 CLIP ViT-L/14 here is the one-off 436 MB download, and once that is cached the page keeps working with the network switched off.
clip vit-l/14clip vit-l/14 onnxclip vit-l/14 browserrun clip vit-l/14 locallyxenova/clip-vit-large-patch14zero-shot image labels modelclip vit-l/14 webgpuclip vit-l/14 download sizefree zero-shot image labels modeltransformers.js model
About 436 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/clip-vit-large-patch14 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.