Platforms · MCPENROUZ

AI agent for Linux computers and servers through MCP

Connect a workstation or Linux server to a compatible AI client. The agent uses outbound HTTPS to MCPENROUZ, so no inbound MCP port is needed.

Free2 devices · 100 tool calls per hour
Example task
Select Linux dev. Read the README in /srv/projects/demo and show the project’s documented test commands.
Your device · your permissions
01

Workstations and VPS

Install the agent under a suitable Linux user and select the project directory.

02

No separate Node.js

The installer contains a private runtime; use compatible Linux with glibc 2.28+.

03

OS account permissions

Start with files only and enable terminal access later.

Get started

  1. Select a Linux build

    Download x64 or ARM64 on the setup page and verify its checksum.

  2. Run the installer

    Run sh /path/to/installer.sh. Installation uses the current user account.

  3. Connect and verify

    Select directories, approve the device and ask the AI to list files in an allowed folder.

Example task

A prompt for your AI
Select Linux dev. Read the README in /srv/projects/demo and show the project’s documented test commands.

Permissions

Root is not required for ordinary installation. An enabled shell inherits the agent user permissions, including groups and sudo capabilities.

Details

Headless server

For a server without a browser, use the mcpenrouz remote command and approve it with PENROUZ ID on another device.

Task completion

For long tasks, use managed processes: start, read output and check completion.

Frequently asked questions

Do I need an inbound server port?

The agent connects to MCPENROUZ over outbound HTTPS. No inbound MCP port is required.

Does it work on Alpine Linux?

Current Linux installers target glibc. These builds do not claim musl/Alpine compatibility.

Which build is verified?

Linux x64 has installation verification. ARM64 remains a preview until checked on matching hardware.

Next steps

MCPENROUZ · FREE

Connect a device. Start with reads.

Choose a working folder and AI client. Add writes and terminal access as your task requires.