Tavily
Connects to the Tavily API for AI-powered web search and content extraction. An agent can run research-oriented searches with optional AI-generated answers, and extract clean content from web pages.
Add to an agent node's tools array with _node_type: "tool-tavily". See Connector Reference — How connectors are used.
Authentication
| Field | Required | Description |
|---|---|---|
api_key | Yes | Tavily API key, sent as Authorization: Bearer <api_key>. Typically a stored secret. |
{
"name": "tavily",
"_node_type": "tool-tavily",
"api_key": "{{ secret.TAVILY_API_KEY }}"
}
Tools
tavily_search
AI-powered web search using Tavily. Returns relevant results with optional AI-summarized answer.
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
query | string | Yes | Search query |
search_depth | string | No | Search depth: basic or advanced (default basic) |
max_results | number | No | Max results (default 5) |
include_answer | boolean | No | Include AI-generated answer |
include_domains | string | No | Comma-separated domains to include |
exclude_domains | string | No | Comma-separated domains to exclude |
Returns: An optional **Answer:** <ai answer> block (if include_answer is true and an answer was returned), followed by a numbered list of results in the form <n>. **<title>** (score: <relevance score>)\n <url>\n <content>. Returns "No results found." if there is no answer and no results.
tavily_extract
Extract clean content from one or more web pages using Tavily.
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
urls | string | Yes | Comma-separated URLs to extract content from |
Returns: For each URL, a ## <url> heading followed by the extracted raw content, separated by ---, or "No content extracted.".