AgentQL MCP Server

This is a Model Context Protocol (MCP) server that integrates AgentQL's data extraction capabilities.

Features

Tools

• extract-web-data - extract structured data from a given 'url', using 'prompt' as a description of actual data and its fields to extract.

Development

Install dependencies:

npm install

Build the server:

npm run build

For development with auto-rebuild:

npm run watch

If you want to try out development version, you can use the following config instead of the default one:

{
  "mcpServers": {
    "agentql": {
      "command": "/path/to/agentql-mcp/dist/index.js",
      "env": {
        "AGENTQL_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

[!NOTE] Don't forget to remove the default AgentQL MCP server config to not confuse Claude with two similar servers.

Debugging

Since MCP servers communicate over stdio, debugging can be challenging. We recommend using the MCP Inspector, which is available as a package script:

npm run inspector

The Inspector will provide a URL to access debugging tools in your browser.

Installation

To use AgentQL MCP Server to extract data from web pages, you need to install it via npm, get an API key from our Dev Portal, and configure it in your favorite app that supports MCP.

Install the package

npm install -g agentql-mcp

Configure Claude

• Open Claude Desktop Settings via ⌘+, (don't confuse with Claude Account Settings)
• Go to Developer sidebar section
• Click Edit Config and open claude_desktop_config.json file
• Add agentql server inside mcpServers dictionary in the config file
• Restart the app

{
  "mcpServers": {
    "agentql": {
      "command": "npx",
      "args": ["-y", "agentql-mcp"],
      "env": {
        "AGENTQL_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

Read more about MCP configuration in Claude here.

Configure VS Code

For one-click installation, click one of the install buttons below:

Manual Installation

Click the install buttons at the top of this section for the quickest installation method. For manual installation, follow these steps:

Add the following JSON block to your User Settings (JSON) file in VS Code. You can do this by pressing Ctrl + Shift + P and typing Preferences: Open User Settings (JSON).

{
  "mcp": {
    "inputs": [
      {
        "type": "promptString",
        "id": "apiKey",
        "description": "AgentQL API Key",
        "password": true
      }
    ],
    "servers": {
      "agentql": {
        "command": "npx",
        "args": ["-y", "agentql-mcp"],
        "env": {
          "AGENTQL_API_KEY": "${input:apiKey}"
        }
      }
    }
  }
}

Optionally, you can add it to a file called .vscode/mcp.json in your workspace. This will allow you to share the configuration with others.

{
  "inputs": [
    {
      "type": "promptString",
      "id": "apiKey",
      "description": "AgentQL API Key",
      "password": true
    }
  ],
  "servers": {
    "agentql": {
      "command": "npx",
      "args": ["-y", "agentql-mcp"],
      "env": {
        "AGENTQL_API_KEY": "${input:apiKey}"
      }
    }
  }
}

Configure Cursor

• Open Cursor Settings
• Go to MCP > MCP Servers
• Click + Add new MCP Server
• Enter the following:
  • Name: "agentql" (or your preferred name)
  • Type: "command"
  • Command: env AGENTQL_API_KEY=YOUR_API_KEY npx -y agentql-mcp

Read more about MCP configuration in Cursor here.

Configure Windsurf

• Open Windsurf: MCP Configuration Panel
• Click Add custom server+
• Alternatively you can open ~/.codeium/windsurf/mcp_config.json directly
• Add agentql server inside mcpServers dictionary in the config file

{
  "mcpServers": {
    "agentql": {
      "command": "npx",
      "args": ["-y", "agentql-mcp"],
      "env": {
        "AGENTQL_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

Read more about MCP configuration in Windsurf here.

Validate MCP integration

Give your agent a task that will require extracting data from the web. For example:

Extract the list of videos from the page https://www.youtube.com/results?search_query=agentql, every video should have a title, an author name, a number of views and a url to the video. Make sure to exclude ads items. Format this as a markdown table.

[!TIP] In case your agent complains that it can't open urls or load content from the web instead of using AgentQL, try adding "use tools" or "use agentql tool" hint.