Tavily

Web search and content extraction API optimized for AI applications — structured results, configurable depth, and research-grade output.

Search Modes

• Quick Search — fast results for simple queries, minimal token usage
• Deep Search — comprehensive results with full content extraction
• Research — multi-step research with structured output schemas

Content Extraction

Extract clean, structured content from any URL. Returns markdown-formatted text with metadata.

Best Practices

• Use quick search for factual lookups and simple queries
• Use deep search when you need comprehensive coverage of a topic
• Configure output schemas for structured, machine-readable results
• Implement post-filtering to reduce noise in search results
• Cache results by query hash to avoid redundant API calls