MirbadLLM documentation

MirbadLLM gathers the machines you already own, from an old gaming PC to a laptop, a Mac or a DGX Spark, into one place that runs AI models and manages them. Start with Get started to try it on a laptop, then follow the setup guide for your real machines.

Get startedTry it on a laptop, run the master, add machines, call a model.

Setup guideInstall on one machine, add Linux, Windows and macOS nodes, troubleshoot.

Windows nodesAn old gaming PC with an NVIDIA card, running llama.cpp.

Mac nodesApple Silicon with llama.cpp on Metal, or MLX.

Configurationdeploy/.env variables, master settings and agent files.

ArchitectureHow the master, the node agents and the router fit together.

Most guides here are copied from the MirbadLLM repository when the site is built, so they always match the code.