Local Commander is a remote MCP server and agent that lets ChatGPT, Claude, Cursor and other AI clients run tasks on the user's own Windows, macOS and Linux computers.
Install a small agent on each of your computers, then ask an AI chat app to check or fix things on any of them, even from your phone.
Local Commander connects AI assistants to the computers a person or small team already owns. The user installs a small agent on each machine, pairs it with an account, and then makes requests from an AI chat client, including from a phone. Requests such as finding out why a disk is full, checking whether a backup copied everything, or fixing a relative's VPN are routed to the right computer and executed there.
The bridge between the chat client and the agents is a remote MCP server at localcmd.app/mcp with OAuth sign-in. The site lists setup routes for ChatGPT, Claude, Cursor, Claude Code, Codex, Raycast, Gemini CLI, Kiro, Antigravity, Devin, Grok Build and LM Studio, and notes that any client with remote MCP and OAuth support can connect. Because the agent connects outward, web-based assistants can reach a home or office machine without opening a port.
Several design choices address the risks of letting a model run commands. Each request records which computer ran it and finishes as done, failed or unknown; unknown outcomes go to a Needs attention list rather than being retried automatically. Independent jobs run in parallel, while jobs on the same folder queue. Users choose between full access to their user account and access to selected folders only, can block specific commands, and can revoke computers or AI clients. The vendor is explicit that the tool is not a sandbox, that commands and results pass through its service, and that removing access does not stop work already in progress. Its use-case guides recommend starting read-only, using a dedicated user, and taking snapshots before changes.
The agent supports Windows 11 x64, macOS 14 or later on Apple silicon, and Linux with a systemd user session, and requires Node.js 20 or later. A usage dashboard tracks model tokens for Claude Code and Codex sessions.
The Free plan covers five computers with unlimited executions and eight parallel slots. The Team plan costs $15 per month per pack, adding ten computers and 24 parallel slots, paid through PayPal. Compared with desktop agents such as Claude Cowork or computer-use tools, Local Commander focuses on reaching many machines from whatever MCP client the user already prefers.
$0
$15
Ready to get started with Local Commander?
View Pricing Options →We believe in transparent reviews. Here's what Local Commander doesn't handle well:
Pricing and features change often. If anything on this Local Commander page is wrong or missing, tell us and include a link to the source. We review every correction.
I ran a six-week evaluation of 15 AI developer tools across four workloads: a Flask-to-FastAPI refactor, a Next.js dashboard built from scratch, 23 bug-fix issues sampled from SWE-bench Verified, and a Rust-to-Go port of a 600-line CLI. **Rankings reflect what shipped working cod
Hidden gems in the AI agent tooling space — from browser infrastructure to memory platforms to observability tools. These production-ready tools solve real problems that most developers haven't discovered yet.