CanIRun.ai

Downloading a multi-gigabyte model only to find your GPU cannot hold it is a waste of an evening. CanIRun.ai checks that before you start.

It is a browser-based hardware analyzer that tells you whether your machine can run a given local AI model, so you can decide what to download with some evidence.

The site reads your GPU, CPU and RAM through browser APIs such as WebGPU, then gives compatibility scores and VRAM requirements for a large library of models, including Llama, Qwen, DeepSeek and Phi.

Each model has a technical breakdown covering architecture type, context window, and memory needs at different quantization levels, from Q2_K up to F16.

That last part matters most in practice, since a model that fails at F16 may fit comfortably at a lower quantization.

CanIRun.ai suits anyone experimenting with open source large language models on their own hardware.

Nothing needs installing, because the check runs in the browser.