home / mcp / cloud browser mcp server

Cloud Browser MCP Server

Provides browser automation capabilities via MCP, including navigate, evaluate, screenshot, click, fill, and content retrieval.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "clpublic-mcp-server-cloudbrowser": {
      "command": "node",
      "args": [
        "path/to/mcp-server-cloudbrowser/cloudbrowser/dist/index.js"
      ],
      "env": {
        "API_KEY": "<YOUR_BROWSERBASE_API_KEY>"
      }
    }
  }
}

You can run a Cloud Browser MCP Server to remotely control a web browser from your MCP client. It lets you navigate pages, execute scripts, take screenshots, and interact with page elements, enabling automated browser tasks inside your MCP workflow.

How to use

Configure your MCP client to connect to the Cloud Browser MCP Server. You will run the local server process on a machine or container, then point your client at it using the provided configuration. Once the server is running, you will see the browser-control tools available in your client interface, typically accessible via the tools or actions area.

Practical usage patterns include navigating to URLs, evaluating JavaScript in the page context, taking screenshots, clicking elements, filling forms, and extracting content. Each tool is exposed as a function you can call from your MCP client with the required inputs.

The available tools map directly to browser actions: navigate to a URL, evaluate code, get the current URL, take screenshots, click selectors, fill inputs, and extract text content.

How to install

Prerequisites: you need Node.js and npm installed on your machine.

Install dependencies and build the server so you have a runnable JavaScript bundle.

npm install
npm run build

Start the server using one of the provided MCP launch configurations. The server runs locally and communicates with your MCP client via stdio.

Code and configuration examples

{
  "mcpServers": {
    "cloudbrowser_node": {
      "command": "node",
      "args": ["path/to/mcp-server-cloudbrowser/cloudbrowser/dist/index.js"],
      "env": {
        "API_KEY": "<YOUR_BROWSERBASE_API_KEY>"
      },
      "transportType": "stdio"
    }
  }
}

{
  "mcpServers": {
    "cloudbrowser_npx": {
      "command": "cmd",
      "args": ["/c", "npx", "-y", "@browsercloud/mcp-server-cloudbrowser"],
      "env": {
        "API_KEY": "<YOUR_BROWSERBASE_API_KEY>"
      },
      "transportType": "stdio"
    }
  }
}

Available tools

cloudbrowser_navigate

Navigate to a specified URL in the browser. Input: url as a string.

cloudbrowser_evaluate

Execute JavaScript code in the browser context. Input: script as a string.

cloudbrowser_get_current_url

Retrieve the current URL of the active browser page.

cloudbrowser_screenshot

Capture a screenshot. Inputs: name (required), selector (optional), width (optional, default 800), height (optional, default 600).

cloudbrowser_click

Click a page element. Input: selector as a CSS selector string.

cloudbrowser_fill

Fill an input field. Inputs: selector (CSS selector) and value (string).

cloudbrowser_get_text

Extract text content from the page. Input: selector (optional CSS selector).