OpenLoL FAQs

Question 1

What is OpenLoL and what does it do?

Accepted Answer

OpenLoL starts a local Model Context Protocol (MCP) server, enabling AI agents to have direct, sandboxed access to your machine's terminal and file system. It allows them to read/write files, execute commands, and inspect system info strictly within a folder boundary you define.

Question 2

How do I connect an AI client like Cursor to OpenLoL?

Accepted Answer

First, install OpenLoL (e.g., `npm i -g openlol`) and start it by exposing a folder (e.g., `openlol .`). Then, configure your MCP-compatible AI client by adding the generated `http://localhost:3333/mcp?token=<your-token>` URL to its MCP server settings.

Question 3

How does OpenLoL ensure the security of my local machine?

Accepted Answer

OpenLoL enforces strict security by restricting all file and command operations to a specified folder, preventing directory traversal. It uses unique UUIDv4 tokens for authentication on each session and defaults to binding only on `localhost` for private, local access.

Question 4

Does OpenLoL rely on cloud services or external relays?

Accepted Answer

No, OpenLoL operates entirely locally. It's designed to run on your machine without any cloud dependencies or data relays, ensuring that all interactions between your AI agent and local machine remain private and on your network.

Question 5

What types of tasks can AI agents perform with OpenLoL?

Accepted Answer

AI agents can perform a wide range of tasks including running shell commands, managing files (read, write, delete, move, list), creating directories, and inspecting system information such as OS details, memory usage, disk usage, and environment variables – all within the designated sandbox.

OpenLoL

OpenLoL

主要功能

使用案例

主要功能

使用案例