Tavily integrations

Integrate with Tavily using LangChain Python.

Tavily is a search engine specifically designed for AI agents, providing search, extract, crawl, and map APIs so developers can connect their applications to real-time online information. Tavily’s primary mission is to deliver factual and reliable information from trusted sources, enhancing the accuracy and reliability of AI-generated content and reasoning.

The langchain-tavily package exposes Tavily’s Search, Extract, Crawl, and Map endpoints as LangChain tools.

Installation and setup

Install the Tavily integration package for LangChain Python:

pip

pip install langchain-tavily

uv add langchain-tavily

Set up a Tavily API key and set it as an environment variable named TAVILY_API_KEY:

export TAVILY_API_KEY="your-api-key"

Tools

TavilySearch

A search tool that returns real-time, LLM-ready results from Tavily’s Search API.

See a usage example.

from langchain_tavily import TavilySearch

TavilyExtract

A tool that returns the cleaned, parsed content of one or more URLs.

See a usage example.

from langchain_tavily import TavilyExtract

TavilyCrawl

A tool that performs a structured web traversal from a base URL, with optional natural-language instructions and path/domain filters.

See a usage example.

from langchain_tavily import TavilyCrawl

TavilyMap

A tool that discovers the URL structure of a site without extracting page content.

See a usage example.

from langchain_tavily import TavilyMap

Components reference

Class	Abstraction	Import path	Description
TavilySearch	Tool	from langchain_tavily import TavilySearch	Returns search results from the Tavily Search API.
TavilyExtract	Tool	from langchain_tavily import TavilyExtract	Extracts cleaned content from one or more URLs.
TavilyCrawl	Tool	from langchain_tavily import TavilyCrawl	Crawls a site starting from a base URL with depth and breadth controls.
TavilyMap	Tool	from langchain_tavily import TavilyMap	Discovers the URL structure of a site without extracting page content.

---

Connect these docs to Claude, VSCode, and more via MCP for real-time answers.

Edit this page on GitHub or file an issue.