Documentation

Tavily AI Search

The Tavily AI Search component provides advanced semantic search capabilities powered by Tavily's AI search technology. It enables precise and contextually relevant search results across various data sources.

Tavily AI Search Component

Tavily AI Search interface

Component Inputs

  • Tavily API Key: Your API key for accessing Tavily services

    Required for authentication with Tavily's API

  • Search Query: The text query to search for

    The search term to be processed by Tavily's AI

  • Search Depth: Determines the thoroughness of search (default: advanced)

    Options include: basic, advanced, comprehensive

  • Search Topic: Category for the search (default: general)

    Helps focus search results on relevant categories

  • Time Range: Timeframe for results (default: any time)

    Allows limiting results to a specific time range

  • Max Results: Maximum number of results to return (default: 5)

    Controls the quantity of search results

  • Include Images: Toggle to include images in search results

    When enabled, results may include relevant images

  • Include Answer: Toggle to include direct answers in response

    When enabled, attempts to provide a direct answer to the query

Component Outputs

  • Data: Structured search results from Tavily
  • Text: Plain text representation of search results

Use Cases

  • Knowledge Retrieval: Find relevant information on specific topics
  • Research Assistance: Gather information from multiple sources
  • Content Generation: Find reference material for content creation
  • Question Answering: Get direct answers to specific questions
  • Fact Checking: Verify information across multiple sources

Best Practices

  • Use specific, clear search queries for best results
  • Choose appropriate search depth based on your needs
  • Limit max results to improve performance for high-volume searches
  • Include topic categorization for more relevant results
  • Enable "Include Answer" for direct question answering
  • Store and reuse results when possible to reduce API usage