Welcome to the Jules Endpoint Agent! This project provides a set of scripts and tools to turn any machine into a secure, remotely-accessible execution endpoint for an AI agent like Jules. This allows an agent to perform development tasks like cloning repositories, running tests, and executing builds in a real, high-performance environment.
This repository contains several methods for installing the agent, each tailored to a specific need or operating system.
This is extremely important. By installing this agent, you are creating a bridge between the public internet and a shell on your machine. The runner scripts are designed to execute arbitrary commands sent to them.
Please follow these security best practices:
- Run this on a dedicated, sandboxed machine or VM. Do not run it on your primary personal or work machine.
- Always use the secure Cloudflare tunnel. Never expose any ports directly.
- Review the installation scripts before running them to understand what they do.
Please choose the installation method that best fits your operating system and use case. Each directory contains a detailed README.md file with specific instructions.
- Go to:
./linux/ - Use this method to install the agent directly on a modern Linux distribution (e.g., Ubuntu, Debian, Fedora, CentOS).
- Go to:
./macos/ - Use this method to install the agent directly on a macOS machine, ideal for testing Apple-specific projects.
- Go to:
./windows/ - Use this method to install the agent directly on Windows, ideal for testing Windows-native projects.
- Go to:
./docker/ - This is the recommended method for most users on any OS who want a secure, isolated, and high-performance Linux environment. It is the best option for tasks that require GPU acceleration.
- Go to:
./virtualbox/ - Use this method if you are on Windows or macOS and prefer the high-isolation of a full Linux VM over the native or Docker installations.
We welcome contributions from the community! Whether it's reporting a bug, suggesting a feature, or submitting code changes, your help is greatly appreciated.
- Contribution Guidelines: Learn how to submit issues and pull requests.
- Manual Testing Guide: Find out how to test your changes before submitting.