mcp-registry/agentql-mcp

    ==================
      
       /// MCP ///
      /// AGE ///
        
    ==================
        
    [server:online]
    [protocol:ready]

agentql-mcp

by tinyfish-io

Model Context Protocol server that plugs AgentQL’s web-data extraction tool into any MCP-compatible client.

94
19
Enterprise
GitHubWebsiteDocumentationNPM

01

extract-web-data

Extract structured data from a given 'url', using 'prompt' as a description of the data and fields to extract.

Installation

1. Prerequisites
• Node.js ≥ 18 LTS and npm (or pnpm / yarn) installed.
• An AgentQL Cloud account and API token (needed to let the MCP server call AgentQL’s extraction engine).
2. Clone or install
a) As a project dependency:
npm install agentql-mcp --save # or pnpm add agentql-mcp
b) Running directly from source:
git clone https://github.com/tinyfish-io/agentql-mcp.git cd agentql-mcp npm install
3. Configuration
Create a .env file (or export in your shell) with at least:
PORT=8080 # Port the MCP server will listen on AGENTQL_API_KEY=<your-agentql-token> # Optional LOG_LEVEL=info # trace | debug | info | warn | error PLAYWRIGHT_HEADLESS=true # false to debug scraping sessions
4. Database (optional)
If you need persistence for scraped artifacts or conversation state, set:
DATABASE_URL=postgres://user:pass@host:5432/db
The server will automatically run migrations on start-up.
5. Start the server
• When installed via npm script:
npx agentql-mcp start
• From source:
npm run start
The server should output `MCP server listening on :8080`.
6. Reverse-proxy (production)
Behind Nginx or Caddy forward traffic to localhost:8080 and enable HTTPS.

Documentation

License: MIT License
Updated 7/30/2025