Skip to main content
Hands-on notebooks you can run locally or on Colab. Each notebook focuses on a single concept with detailed code and explained outputs.

View all notebooks on GitHub

Getting Started

Search, Extract & Crawl

Core API walkthrough — learn the three foundational endpoints with guided examples.

Web Agent

Build an LLM-powered web research agent that uses Tavily tools end-to-end.

Hybrid Agent

Combine a Chroma vector store with Tavily web search for hybrid answers.

Product News Tracker

Track product updates using domain-filtered and news-specific search parameters.

Research

Polling

Submit async research jobs and poll for completion — ideal for background processing.

Streaming

Stream real-time progress events and answers during research execution.

Structured Output

Get research results in custom JSON schemas with Pydantic-style definitions.

Query Refinement

Multi-turn clarification flow that refines user prompts before running research.

Hybrid Research

Combine Tavily research with your internal data for comprehensive reports.

Crawl

Crawl Fundamentals

Crawl and map basics — depth, breadth, and path controls with visualization.

Crawl to RAG

Full pipeline from website crawl to chunked embeddings to Q&A with a vector store.

Agent Grounding

LangGraph-style research agent that uses search, map, and extract for autonomous web research.

Data Collection

Crawl websites and export pages as organized PDF files for offline analysis.

Integrations

LangChain

LangChain tool wrappers for Search, Extract, Map, and Crawl with filtered and news variants.

Remote MCP

Configure Tavily’s remote MCP server and call it from an OpenAI client.

AWS Strands

Deep research agent using AWS Strands with Bedrock and Tavily search, extract, and crawl tools.