Built by Metorial, the integration platform for agentic AI.
Traverse a website starting from a root URL, following links across pages with built-in content extraction. Returns extracted content from each discovered page. Supports natural language instructions to guide the crawler, path/domain filtering via regex, and configurable depth and breadth controls.
Conduct in-depth, multi-step research on a topic. Tavily autonomously performs multiple searches, analyzes sources, and produces a comprehensive research report with citations. The task is queued asynchronously - this tool creates the research task and polls until it completes or times out.
Perform an AI-optimized web search that returns relevance-scored results with content snippets. Supports configurable search depth, topic categories (general, news, finance), time filtering, domain filtering, country boosting, and optional LLM-generated answers. Can also return images and full raw page content.
Extract clean, structured content from one or more URLs. Returns full page content in markdown or plain text format. Supports query-based reranking to surface the most relevant content chunks, and optional image extraction.
Retrieve API credit usage information for your Tavily account. Returns a breakdown of credits consumed by each endpoint (search, extract, crawl, map, research) along with plan limits and pay-as-you-go usage.
Discover and list all URLs on a website by traversing its link structure, without extracting content. Returns a flat list of discovered URLs. Useful for understanding site structure before crawling or extracting specific pages.