Tavily

Tavily is a real-time search and web data retrieval API for developers and teams building AI agents, RAG pipelines, and knowledge-based assistants that need fresh web context.

Key capabilities include:

• Search API that combines web search and scraping, returning clean, structured results optimized for LLM consumption
• Extract API for content scraping from up to 20 URLs per call, designed for reliable ingestion into retrieval systems
• Crawl API for domain mapping and crawling to collect and organize web content at scale
• Integrations for agent and RAG frameworks such as LangChain and LlamaIndex, plus controls such as include_raw_content and max_tokens_per_page

Pricing: Freemium with a free tier (1,000 API credits/month), pay-as-you-go at $0.008/credit, monthly plans at $0.0075 to $0.005 per credit, and enterprise options (custom). Paid plans support higher rate limits; email support is available.

Notable for its focus on LLM-ready retrieval and secure web access patterns for agents, plus endpoints that cover search, extraction, and crawling through a single API layer.