tavily-python

Why we included this project

Agents and RAG pipelines that need to ground answers in current web content usually end up gluing together a search API and a scraper. This SDK saves you that wiring: it wraps Tavily's hosted search in a small Python client with sync and async variants, so search, extraction, and the heavier crawl and research endpoints fit into an existing codebase without hand-rolled HTTP calls. The search method returns structured results with configurable depth, domain allow and deny lists, and optional LLM-generated answers, which makes it simple to feed clean snippets into a prompt or vector store. A keyless mode covers search and extraction for quick prototyping, and the remaining endpoints open up once you add an API key. Teams that would rather use a managed search backend than run their own crawler infrastructure get a practical integration point here.