rod
Model Context Protocol Server of Rod
claude mcp add --transport stdio go-rod-rod-mcp rod-mcp -c rod-mcp.yaml
How to use
The Rod MCP server provides browser automation capabilities driven by the Go-based Rod library. It exposes a set of MCP tools that let language models or automated agents interact with web pages—for example, clicking elements, capturing screenshots, generating PDFs, and performing other common browser automation tasks. The server is designed to be headless-friendly and supports options for running with or without a GUI, as well as proxy configuration and browser customization. You can configure the server through a YAML/JSON configuration file and reference the rod-mcp executable via the provided command and arguments.
To use it, start the Rod MCP server using the configured command, typically by launching rod-mcp with a configuration file (rod-mcp.yaml). The server then exposes endpoints and tools that your LLM or automation client can invoke to perform browser actions, retrieve results (e.g., HTML content, element states, or screenshots), and chain operations together in a workflow.
How to install
Prerequisites:
- Go 1.23 or higher
- Chrome/Chromium browser
Installation steps:
-
Clone the repository: git clone https://github.com/go-rod/rod-mcp.git cd rod-mcp
-
Install dependencies and build: go mod tidy go build
-
Prepare a configuration file (rod-mcp.yaml) with your desired settings (see Configuration section in the README).
-
Run the server (example): ./rod-mcp -c rod-mcp.yaml
Notes:
- Ensure that the browser executable path is correctly set if you’re not using the system default (see browserBinPath in the configuration).
- If you want to run in headless mode or with a proxy, configure the corresponding options in rod-mcp.yaml before starting.
Additional notes
Configuration tips:
- serverName: Change the server display name from the default to something meaningful for your environment.
- browserBinPath: Provide an explicit path to the Chrome/Chromium binary if needed.
- headless: Set to true to run in headless mode; false may show a GUI depending on your environment.
- browserTempDir: Location for temporary browser data; ensure the directory is writable.
- noSandbox: Disable sandboxing if you’re running in environments where sandbox is restricted.
- proxy: Configure a proxy (including SOCKS5) if your network requires it.
- rod-mcp.yaml: The configuration file referenced by -c; you can customize available options and defaults therein.
Common issues:
- Mismatched Chrome/Rod versions leading to incompatibilities – ensure compatible versions.
- Insufficient permissions for browser temp directory or config file paths.
- Network/firewall rules blocking proxy or endpoints used by automation tasks.
Related MCP Servers
ollama
An MCP Server for Ollama
time
⏰ Time MCP Server: Giving LLMs Time Awareness Capabilities
workflowy
Powerful CLI and MCP server for WorkFlowy: reports, search/replace, backup support, and AI integration (Claude, LLMs)
browserai
A powerful Model Context Protocol (MCP) server that provides an access to serverless browser for AI agents and apps
gtm
An MCP server for Google Tag Manager. Connect it to your LLM, authenticate once, and start managing GTM through natural language.
mcp-human
Human Assistance for AI Assistants