LLMule

Run a chat or coding prompt against an open model on your own machine, or hand it to another person's hardware on the network, without sending your data to a central provider.

That is the basic idea behind LLMule, a decentralized AI tool for local and peer-to-peer model execution.

It is an open-source project with a community-driven library of models, and it works with multiple AI frameworks.

If you share your own compute power, you can earn credits in return.

The privacy pitch is simple: data stays local.

The catches are real, though.

Which models are available depends on what the community contributes, and speed will vary with your hardware and network connection.

Setting up and managing a P2P network may take some effort, and the software is in alpha or beta, so expect bugs and some instability.