Selenium MCP (mcp-selenium)

Add to your client

Copy the config for your MCP client and paste it into its config file.

{
  "mcpServers": {
    "selenium-mcp-mcp-selenium": {
      "command": "npx",
      "args": [
        "-y",
        "@angiejones/mcp-selenium"
      ]
    }
  }
}

Before you start

• Node.js with npm (the server runs via npx or a global install)
• A supported browser installed locally: Chrome, Firefox, Edge, or Safari
• Matching WebDriver where needed (e.g. chromedriver). Safari requires macOS, a one-time `sudo safaridriver --enable`, and 'Allow Remote Automation' turned on; Safari has no headless mode
• No API key or account — authentication is none

About Selenium MCP (mcp-selenium)

Tools & capabilities (18)

When to use it

• Use it when you want an agent to run local end-to-end browser tests without paying for a cloud browser.
• Use it when you need cross-browser coverage — Chrome, Firefox, Edge, and Safari on your own machine.
• Use it when a task requires real element interaction: clicking, typing, file uploads, frames, and alerts.
• Use it when you want full control with no API key, account, or per-request billing.
• Use it when you need to inspect the accessibility tree or BiDi diagnostics during automation.

Quick setup

1. 1Ensure Node.js and a supported browser (plus its driver, e.g. chromedriver) are installed.
2. 2Add the server to your client — e.g. `claude mcp add selenium -- npx -y @angiejones/mcp-selenium@latest`, or configure command "npx" with args ["-y", "@angiejones/mcp-selenium@latest"] in Cursor/Windsurf.
3. 3For Safari, run `sudo safaridriver --enable` once and enable 'Allow Remote Automation'.
4. 4Restart your MCP client.
5. 5Verify by asking the client to start_browser and navigate to a URL.

Selenium MCP (mcp-selenium) FAQ