How to add AgentQL MCP to Claude Desktop

Turn any web page into structured data — AgentQL's prompt-driven extraction as a single MCP tool. Paste the config into ~/Library/Application Support/Claude/claude_desktop_config.json and restart Claude Desktop.

Last updated June 15, 2026 · 400★ · stdio · apikey · official

Claude Desktop config for AgentQL MCP

npx -y agentql-mcp

{
  "mcpServers": {
    "agentql-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "agentql-mcp"
      ],
      "env": {
        "AGENTQL_API_KEY": "<your-agentql-api-key>"
      }
    }
  }
}

Setup steps

1Open Claude Desktop → Settings → Developer → Edit Config (this opens ~/Library/Application Support/Claude/claude_desktop_config.json).
2Paste the AgentQL MCP config below under the top-level "mcpServers" key.
3Fill in any placeholder secrets (API keys, paths) in the snippet.
4Save the file, then fully quit and reopen Claude Desktop.
5Open a chat and confirm AgentQL MCP's tools appear under the 🔌 tools menu.

Before you start

Node.js with npm (run via `npx -y agentql-mcp` or install globally)
An AgentQL API key from the AgentQL Dev Portal (https://dev.agentql.com) set as AGENTQL_API_KEY
An MCP-compatible client (Claude Desktop, Cursor, VS Code, or Windsurf)

What AgentQL MCP can do in Claude Desktop

extract-web-data

Extract structured data from a given URL, using a natural-language prompt to describe the fields to extract.

Security

Needs an AGENTQL_API_KEY that bills against your AgentQL account; store it in env and rotate if leaked. Pages are fetched and parsed through AgentQL's service, and extracted data flows back to your model as untrusted content.

AgentQL MCP + Claude Desktop FAQ

Where is the Claude Desktop config file?

Claude Desktop reads MCP servers from ~/Library/Application Support/Claude/claude_desktop_config.json. Paste the AgentQL MCP config there under the "mcpServers" key and restart the client.

Is AgentQL MCP safe to use with Claude Desktop?

How do I authenticate?

Create an API key in the AgentQL Dev Portal and pass it as the AGENTQL_API_KEY environment variable in your MCP client's server config.

What tools does it expose?

Just one: `extract-web-data`, which takes a `url` and a `prompt` describing the data and fields to extract, and returns structured data.

The agent isn't calling AgentQL — what do I do?

Add an explicit hint to your prompt such as "use tools" or "use the agentql tool" so the model invokes the extraction tool.

View repo Full AgentQL MCP page