LOCAL AI · INTELLIGENCE AND COMPLETE

Intelligence on your device.

Jan runs Qwen3 language models locally. Ask questions, draft text, and discuss documents on your own computer after setup. The included models offer different memory and performance tradeoffs.

GRIDGONE Local AI screen with Jan local server connection settings — Local AI connects to Jan on your computer Actual dashboard · features vary by edition

Set up locally.

Install Jan, import a supplied model, and start its local API server to use the dashboard’s AI screen. This marketing website does not run local AI or send prompts to Jan. The offline dashboard connects to the local server you configure.

Check your hardware.

Allow at least 8 GB RAM; 16 GB or more is recommended for larger models. Jan’s local inference requires a supported Windows machine or Apple Silicon Mac. Local inference on Intel Macs is not supported. Model loading and performance on physical hardware have not yet been verified for these packages.

AI can give incorrect answers. Review its output, especially for medical, safety, or other consequential decisions.

Compatibility details →