browsercat

BrowserCat MCP Server provides cloud-based browser automation through the BrowserCat service. It exposes tools that let an LLM drive a real browser to navigate pages, take screenshots, click and interact with elements, fill forms, select options, and run JavaScript within the page. You can retrieve browser console logs and capture screenshots by name, enabling rich web automation workflows without local browser installations. Use the available tools to compose sequences such as navigating to a URL, logging results, interacting with page elements, and validating page content via script evaluations. The server is designed to be accessible via the MCP protocol, so you can invoke actions through your preferred MCP-enabled client and capture outputs like console logs or images as needed.

Prerequisites:

• Node.js (and npm or npx) installed on your development machine
• A BrowserCat API key (required) at https://browsercat.xyz/mcp

Installation steps:

1. Obtain your BrowserCat API key and keep it handy.

2. Ensure Node.js is installed by running node -v and npm -v.

3. You don’t need a local install for the server when using NPX, but you can verify the setup by running a quick NPX command:

   npx -y @browsercatco/mcp-server

4. Run the MCP server with your API key set in the environment. Example (Unix-like shell):

   export BROWSERCAT_API_KEY=your-api-key-here npx -y @browsercatco/mcp-server

5. If you prefer, you can integrate this into a script or a Docker-based workflow once you’re comfortable with the environment variable configuration.

Note: The environment variable BROWSERCAT_API_KEY is required for authentication with BrowserCat.

Tips and considerations:

• Always provide your BROWSERCAT_API_KEY in the environment before starting the server. Without it, browser automation requests will be rejected.
• The tools (navigate, screenshot, click, hover, fill, select, evaluate) map to browser interactions; construct your requests with the appropriate input fields as described in the server's capabilities.
• Screenshots are accessible by name via the screenshot resource (screenshot://<name>).
• Console logs will stream browser console output to console://logs for debugging.
• If you encounter rate limits or authentication errors, check your BrowserCat plan and API key permissions.
• This configuration uses NPX to run the server; you can switch to a persistent deployment method later if needed and keep the same environment variable approach.