Running large AI models locally in the browser requires specific hardware and software features. The most critical enabler is WebGPU, which allows browsers to offload compute-intensive tasks to the GPU. This tool checks if your browser exposes the WebGPU API and estimates your available processing power based on hardware concurrency and device memory.
How It Works
We look for the navigator.gpu object to determine WebGPU support. If available, we query the default adapter to see if it supports compute shaders. We also check navigator.hardwareConcurrency for CPU cores and estimate memory limits, as local AI models like Llama 3.2 or Whisper require significant RAM and VRAM.
What The Results Mean
A 'Yes' for WebGPU doesn't guarantee smooth performance, but it is the primary gatekeeper for high-speed local inference. If WebGPU is missing, you might still run smaller models via WebAssembly (WASM) on the CPU, but performance will be slower.
How this check works
Open this page and the embedded script queries only documented browser APIs. Nothing is installed, no account is needed, and results reflect your current browser profile — including extensions, private mode and OS display scaling.
Limitations you should know about
Support varies by browser. Privacy features (tracking prevention, user-agent reduction, fingerprint randomization) can generalize or withhold values. Treat results as a best-effort live snapshot rather than a certified hardware inventory.
Privacy
Detection runs in your browser. We do not store fingerprints or device reports. The only network call any tool makes is the optional public-IP lookup on /my-ip, disclosed in the privacy policy.