Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.tavily.com/llms.txt

Use this file to discover all available pages before exploring further.

​
Introduction

The Tavily MCP Server connects your ADK agent to Tavily’s AI-focused search, extraction, and crawling platform. This gives your agent the ability to perform real-time web searches, intelligently extract specific data from web pages, and crawl or create structured maps of websites.

​
Prerequisites

Before you begin, make sure you have:
  • Python 3.9 or later
  • pip for installing packages
  • A Tavily API key (sign up for free if you don’t have one)
  • A Gemini API key for Google AI Studio

​
Installation

Install ADK by running: 
pip install google-adk mcp

​
Building Your Agent

​
Step 1: Create an Agent Project

Run the adk create command to start a new agent project: 
adk create my_agent
This creates a new directory with the following structure: 
my_agent/
    agent.py      # main agent code
    .env          # API keys or project IDs
    __init__.py

​
Step 2: Update Your Agent Code

Edit the my_agent/agent.py file to integrate Tavily. Choose either Remote MCP Server or Local MCP Server: 
from google.adk.agents import Agent
from google.adk.tools.mcp_tool.mcp_session_manager import StreamableHTTPServerParams
from google.adk.tools.mcp_tool.mcp_toolset import MCPToolset
import os

# Get API key from environment
TAVILY_API_KEY = os.getenv("TAVILY_API_KEY")

root_agent = Agent(
    model="gemini-2.5-pro",
    name="tavily_agent",
    instruction="You are a helpful assistant that uses Tavily to search the web, extract content, and explore websites. Use Tavily's tools to provide up-to-date information to users.",
    tools=[
        MCPToolset(
            connection_params=StreamableHTTPServerParams(
                url="https://mcp.tavily.com/mcp/",
                headers={
                    "Authorization": f"Bearer {TAVILY_API_KEY}",
                },
            ),
        )
    ],
)

​
Step 3: Set Your API Keys

Update the my_agent/.env file with your API keys: 
echo 'GOOGLE_API_KEY="YOUR_GEMINI_API_KEY"' >> my_agent/.env
echo 'TAVILY_API_KEY="YOUR_TAVILY_API_KEY"' >> my_agent/.env
Or manually edit the .env file: 
GOOGLE_API_KEY="your_gemini_api_key_here"
TAVILY_API_KEY="your_tavily_api_key_here"

​
Step 4: Run Your Agent

You can run your ADK agent in two ways:

​
Run with Command-Line Interface

Run your agent using the adk run command: 
adk run my_agent
This starts an interactive command-line interface where you can chat with your agent and test Tavily’s capabilities.

​
Run with Web Interface

Start the ADK web interface for a visual testing experience: 
adk web --port 8000
Note: Run this command from the parent directory that contains your my_agent/ folder. For example, if your agent is inside agents/my_agent/, run adk web from the agents/ directory. This starts a web server with a chat interface. Access it at http://localhost:8000, select your agent from the dropdown, and start chatting.

​
Example Usage

Once your agent is set up and running, you can interact with it through the command-line interface or web interface. Here’s a simple example: User Query: 
Find all documentation pages on tavily.com and provide instructions on how to get started with Tavily
The agent automatically combines multiple Tavily tools to provide comprehensive answers, making it easy to explore websites and gather information without manual navigation. Tavily-ADK

​
Available Tools

Once connected, your agent gains access to Tavily’s powerful web intelligence tools: Execute a search query to find relevant information across the web.

​
tavily-extract

Extract structured data from any web page. Extract text, links, and images from single pages or batch process multiple URLs efficiently.

​
tavily-map

Traverses websites like a graph and can explore hundreds of paths in parallel with intelligent discovery to generate comprehensive site maps.

​
tavily-crawl

Traversal tool that can explore hundreds of paths in parallel with built-in extraction and intelligent discovery.