# WECOMMANDER

> WECOMMANDER is a remote MCP (Model Context Protocol) server that lets AI assistants such as Claude, ChatGPT and Cursor use the terminals and files of your own Windows, macOS and Linux machines through one MCP URL, with human approval for risky actions, per-device limits and a full audit log.

Key facts:

- One MCP URL for all your machines: https://wecommander.app/mcp (OAuth sign-in from Claude, ChatGPT and other MCP clients).
- A small agent runs on each machine (Windows, macOS, Linux) and keeps one outbound encrypted connection; no inbound ports, VPN or SSH keys are needed.
- The agent runs as a normal user (not root/SYSTEM) and only inside the folders you allow.
- High-risk actions (running commands, deleting or moving files, killing processes, opening terminals) wait for a human approval of that exact action, unless you allow a command pattern or switch a device to autonomous mode for a set time.
- Approvals arrive in the web dashboard, by e-mail and in the Management Center desktop app as Windows notifications with Approve / Deny buttons.
- Secrets in command and terminal output (tokens, passwords, keys) are masked on the device before the result reaches the AI.
- Every action is recorded in an audit log; file edits keep undo snapshots.
- Tools: read, write, edit, search and move files; run commands; long-running background tasks; interactive terminals; process lists; device health snapshots; log search; PDF, Word and Excel reading.
- Local mode: the same agent can serve MCP over stdio to a local client (Claude Desktop, Cursor) with no account or server.

## Product

- [Your AI, on every machine](https://wecommander.app/): Connect Claude, ChatGPT or any MCP client to the terminals and files of your Windows, macOS and Linux machines through one MCP URL, with human approvals, a non-root agent and a full audit trail.
- [Features](https://wecommander.app/features): Terminals, files, background tasks, logs and health of your machines for AI assistants, with approvals, allowed folders, secret masking, undo and an audit log.
- [How it works](https://wecommander.app/how-it-works): An agent on each machine keeps an outbound encrypted connection to the WECOMMANDER gateway; AI clients connect to one MCP URL; risky actions wait for your approval.
- [Pricing](https://wecommander.app/pricing): WECOMMANDER plans for individuals and teams. See the current prices and limits on this page.

## Docs

- [Install](https://wecommander.app/install): Install the WECOMMANDER agent on Windows, macOS or Linux with one command and connect it to your AI client.
- [Documentation](https://wecommander.app/docs): Set up WECOMMANDER with Claude, ChatGPT, Cursor and other MCP clients, run the agent in local mode, and manage approvals.
- [Security model](https://wecommander.app/security): How WECOMMANDER keeps AI access to your machines under control: outbound-only agent, allowed folders, approvals, secret masking, audit log.
- [Frequently asked questions](https://wecommander.app/faq): Answers about connecting AI assistants to your computers and servers with WECOMMANDER: safety, supported systems, clients and local mode.
- [Ways to give an AI access to your machines](https://wecommander.app/alternatives): Ways to give an AI assistant access to your computers and servers: local MCP file servers, SSH-based tools, remote management tools and WECOMMANDER, compared.

## Guides

- [Give Claude secure access to your server's terminal](https://wecommander.app/guides/claude-server-terminal): Step by step: let Claude run commands and edit files on your Linux, Windows or macOS server without opening SSH, with approval of each command.
- [What is a remote MCP server?](https://wecommander.app/guides/remote-mcp-server): A remote MCP server lets AI assistants use tools over the internet with one URL and OAuth sign-in. When you need one and how WECOMMANDER uses it to reach your machines.
- [Let ChatGPT work with files on your Windows PC](https://wecommander.app/guides/chatgpt-windows-files): Connect ChatGPT to a folder on your Windows computer to read, search and edit files, with your approval for anything risky.
- [Keep a human in the loop: approving AI actions](https://wecommander.app/guides/ai-approvals): Keep a human in the loop when an AI assistant works on your computers: single-use approvals, allowed commands, trust levels and autonomous mode with automatic expiry.

## Optional

- [Full text of the public pages](https://wecommander.app/llms-full.txt)
- [Privacy](https://wecommander.app/privacy)
- [Terms](https://wecommander.app/terms)