A Tool to Check If Your Computer Can Run Local AI Models
CanIRun.ai, a web-based utility, tells users whether their computer's hardware can run popular open-source AI models locally. The site recently gained attention after being highlighted in a Hacker News post.
CanIRun.ai attempts to automatically detect a visitor's GPU, available VRAM, and system RAM, then compares that against a database of open-source models — the site references options such as Meta's Llama 3, Microsoft's Phi-3, and various Mistral releases. It reports which models a given machine can run and, in many cases, which quantized (compressed) versions are likely to perform best on that hardware.
The tool addresses a real pain point for people trying to run AI models locally rather than through cloud APIs: figuring out VRAM requirements, quantization levels, and processor capability has typically required manual research across forums and documentation. By automating hardware detection and matching it to model requirements, CanIRun.ai removes some of that guesswork for newcomers to local inference.
The tool is free to use directly from a browser, with no indication from the site of account requirements or paid tiers.