What is RunOnMine?

A security boundary between AI and your computer.

RunOnMine lets an AI assistant use useful tools on a machine you own while keeping execution, policy, credentials, approvals and audit decisions local.

The short answer

RunOnMine is a local-first Model Context Protocol (MCP) gateway and desktop control center. It sits between an AI client and machine tools, then decides locally whether each requested action is allowed, needs approval or must be denied.

What AI can use

RunOnMine can expose selected-root file tools, terminal and process tools, isolated browser automation, desktop actions and platform-native operations. Availability remains constrained by the local operating system and configured policy.

What stays local

Selected filesystem roots, credentials, policy evaluation, exact-action approvals, audit state and execution live on the machine being controlled. The MCP listener is kept on loopback; supported remote connectors connect outward.

Who it is for

Developers and machine owners who want AI to do real work on a local workstation, development machine or server without treating a raw shell, SSH account or public MCP endpoint as the security model.