> ## Documentation Index
> Fetch the complete documentation index at: https://docs.shanone.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Tavily

> AI-optimized web search, extraction, crawling, and research — built for agent workflows

# Tavily

Shanone's Tavily integration gives your agent 5 tools for AI-optimized web search and content retrieval — search results are pre-cleaned and ranked for LLM consumption rather than raw HTML.

## Getting Started

<Steps>
  <Step title="Get a Tavily API key">
    Sign up at [tavily.com](https://tavily.com) and copy an API key from your Tavily dashboard.
  </Step>

  <Step title="Connect in Shanone">
    Ask your agent to search the web with Tavily, or add the API key proactively from **Integrations** in the Shanone dashboard.
  </Step>

  <Step title="Retry your request">
    Once the key is stored, `shanone_execute_tool` calls for `tavily_*` tools will succeed.
  </Step>
</Steps>

<Note>
  Tavily uses API key authentication rather than OAuth. If no key is connected for your account, Shanone falls back to a shared server-side key where configured — check with your admin if `tavily_*` calls behave unexpectedly.
</Note>

## Available Tools

Shanone provides **5 tools** for Tavily:

<AccordionGroup>
  <Accordion title="Search">
    `tavily_search` — AI-optimized web search with configurable depth, topic filters, and result count
  </Accordion>

  <Accordion title="Extract">
    `tavily_extract` — pull clean, structured content from one or more URLs
  </Accordion>

  <Accordion title="Crawl">
    `tavily_crawl` — recursively crawl a site starting from a base URL
  </Accordion>

  <Accordion title="Map">
    `tavily_map` — discover a site's URL structure without extracting full content
  </Accordion>

  <Accordion title="Research">
    `tavily_research` — multi-step, agentic research that combines search and extraction into a single synthesized answer
  </Accordion>
</AccordionGroup>

## Common Use Cases

<CardGroup cols={2}>
  <Card title="Competitive research" icon="magnifying-glass">
    Search and extract pricing/feature pages from competitor sites for a comparison doc
  </Card>

  <Card title="Fact-checking" icon="circle-check">
    Verify a claim in a draft by searching for corroborating sources
  </Card>

  <Card title="Site discovery" icon="sitemap">
    Map a documentation site's structure before deciding which pages to crawl
  </Card>

  <Card title="Deep research reports" icon="file-text">
    Use `tavily_research` to produce a synthesized, cited answer to an open-ended question
  </Card>
</CardGroup>

## Troubleshooting

<AccordionGroup>
  <Accordion title="tavily_search returns stale or irrelevant results">
    Narrow the query and set an explicit `topic` (e.g. `"news"`) or a `time_range` if the tool schema supports it — check `shanone_get_tool_schema` for `tavily_search` to see available filters.
  </Accordion>

  <Accordion title="tavily_crawl times out or returns partial results">
    Crawls of large sites can exceed reasonable single-call latency. Scope the crawl with a narrower base URL or lower depth/limit parameter, or use `tavily_map` first to plan a targeted crawl.
  </Accordion>

  <Accordion title="401/403 errors on any tavily_* call">
    Your Tavily API key is missing, invalid, or has hit its plan's request quota. Re-check the key under **Integrations** in the Shanone dashboard against your Tavily account.
  </Accordion>
</AccordionGroup>