Agent Development Kit (ADK) Integration

AgentQL integrates with Google's Agent Development Kit (ADK), allowing you to build powerful agents to navigate the web and extract data from any website.

Agent Development Kit (ADK) is an open-source framework from Google for building and deploying AI agents that can interact with the web, tools, and other APIs using the Model Context Protocol (MCP).

Set up AgentQL with ADK

Before you start, get your Google API key and your AgentQL API key.

1. Install the Agent Development Kit:

pip install google-adk

2. Create a new agent project:

adk create agentql_agent
cd agentql_agent

3. Configure the environment variables by adding your API keys to the .env file in your project:

GOOGLE_API_KEY=<YOUR_GOOGLE_API_KEY>
AGENTQL_API_KEY=<YOUR_AGENTQL_API_KEY>

4. Set up your agent with AgentQL in your agent.py file:

import os
from google.adk.agents import Agent
from google.adk.tools.mcp_tool.mcp_session_manager import StdioConnectionParams
from google.adk.tools.mcp_tool.mcp_toolset import MCPToolset
from mcp import StdioServerParameters

AGENTQL_API_KEY = os.getenv("AGENTQL_API_KEY")

root_agent = Agent(
    model="gemini-2.5-pro",
    name="agentql_agent",
    instruction="Help users get information from AgentQL",
    tools=[
        MCPToolset(
            connection_params=StdioConnectionParams(
                server_params = StdioServerParameters(
                    command="npx",
                    args=[
                        "-y",
                        "agentql-mcp",
                    ],
                    env={
                        "AGENTQL_API_KEY": AGENTQL_API_KEY,
                    }
                ),
                timeout=300,
            ),
        )
    ],
)

Usage

Running your agent

You can run your agent using either of the following commands:

Command-line interface:

adk run agentql_agent

Web interface:

adk web --port 8000

Example prompts

Once your agent is running, use this prompt to test it out:

Return the top 20 news articles from https://news.ycombinator.com/

Here are some other examples:

• 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.
• Get this recipe for me: https://www.justonecookbook.com/pressure-cooker-japanese-curry/

Available tools

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

Your agent will automatically have access to these tools and can use them as needed to complete tasks.

Support

If you have any questions about using AgentQL with ADK:

• Join the AgentQL Discord community
• Check out the ADK documentation
• View the AgentQL MCP server on GitHub

Examples

Here are some example use cases for AgentQL with ADK:

• Market Research Agent - Extract product listings, prices, and reviews from e-commerce sites
• News Aggregator - Collect and summarize articles from multiple news sources
• Job Search Assistant - Scrape job postings and match them to user preferences
• Recipe Collector - Extract recipe details including ingredients and instructions from cooking websites