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

# Tavily

> Connect Tavily to your Atako agents — 5 read and 0 write actions.

Let your agents search the web, extract content from URLs, map and crawl sites via the Tavily API.

## Connection

* **Authentication**: API key.

<Note>
  Sign in at app.tavily.com → Overview page, "API Keys" section → Create API key. Name it and copy the key (starts with tvly-). Every call consumes credits billed by Tavily on your account.

  See [Tavily's documentation](https://app.tavily.com).
</Note>

## Read actions (5)

| Action      | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
| ----------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `crawl`     | Crawl a website and return extracted content from all discovered pages. Consumes 1 credit per 10 successful pages (2 credits if instructions provided). Arguments: url (string, required — the root URL to crawl), instructions (string, optional — natural-language instructions for the crawler), max\_depth (integer 1–5, optional — default 1), max\_breadth (integer 1–500, optional — default 20), limit (integer ≥1, optional — total pages to process, default 50), select\_paths (array of strings, optional — regex patterns to include), exclude\_paths (array of strings, optional — regex patterns to exclude), allow\_external (boolean, optional — include external links, default true), extract\_depth ("basic" \| "advanced", optional — default "basic"), format ("markdown" \| "text", optional — default "markdown"). |
| `extract`   | Extract content from one URL or batch up to 20 URLs. Cost: basic — 1 credit per 5 successful URLs; advanced — 2 credits per 5 successful URLs. Arguments: urls (string or array of 1–20 strings, required — URLs to extract), query (string, optional — rerank chunks by relevance to this query), extract\_depth ("basic" \| "advanced", optional — default "basic"; advanced costs 2 credits per 5 URLs), format ("markdown" \| "text", optional — default "markdown").                                                                                                                                                                                                                                                                                                                                                                  |
| `get_usage` | Get the API key and account usage (credits used, plan limits). Free — consumes no credits.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
| `map`       | Map a website and return all discovered URLs. Consumes 1 credit per 10 successful pages (2 credits if instructions provided). Arguments: url (string, required — the root URL to map), instructions (string, optional — natural-language instructions for the crawler), max\_depth (integer 1–5, optional — default 1), max\_breadth (integer 1–500, optional — default 20), limit (integer ≥1, optional — total pages to process, default 50), select\_paths (array of strings, optional — regex patterns to include), exclude\_paths (array of strings, optional — regex patterns to exclude), allow\_external (boolean, optional — include external links, default true).                                                                                                                                                               |
| `search`    | Search the web with Tavily. Consumes 1 credit (basic) or 2 credits (advanced). Arguments: query (string, required — the search query), search\_depth ("basic" \| "advanced" \| "fast" \| "ultra-fast", optional — controls latency vs relevance, default "basic"; advanced costs 2 credits), topic ("general" \| "news" \| "finance", optional — default "general"), max\_results (integer 0–20, optional — max results to return, default 10), time\_range ("day" \| "week" \| "month" \| "year", optional — filter by publish date), include\_answer (boolean \| "basic" \| "advanced", optional — include an LLM-generated answer), include\_domains (array of strings, optional — restrict to these domains, max 300), exclude\_domains (array of strings, optional — exclude these domains, max 150).                                 |

## Write actions (0)

This connector is read-only — it exposes no write actions.

## Permissions

Every action above must be explicitly granted to an agent before it can be used. See [Permissions](/integrations/permissions) for the grant model and [Security](/integrations/security) for how credentials are protected.