tavily-search

To get LLM-ready web search results, use tavily-search to retrieve current information with relevance scores and content snippets formatted as JSON. This provides structured metadata tailored for language models without noisy raw search data.

Yes, you can filter web search results by specific domains and time ranges. The tavily-search skill supports domain inclusion or exclusion, time range filters, and topic filters to retrieve domain-specific pages and recent news articles.

Yes, you need the Tavily CLI (tvly) installed to retrieve current news articles. The tavily-search skill requires this command-line tool to execute queries and return JSON output with snippets and relevance scores.

The best way to extract full page content during a web search is using tavily-search with its full-content extraction option. This includes raw page content in the results, avoiding a separate extraction step for your LLM workflows.

You can control the depth and result limit for multi-depth web queries using tavily-search options. It supports configuring depth levels and max-results limits to customize the amount and detail of information retrieved from the web.