extract
Extract and process content from a list of URLs using Tavily API. Can handle up to 20 URLs at once.
LABSCO SUMMARY
The choice of search mode is made in the tool name, not in a parameter.
search, searchContext and searchQNA stand for different intents — results, context, an answer — so a model picks the shape it wants rather than tuning options on one endpoint. extract is the other half: hand it URLs you already hold and it returns their processed content. Each call carries an options object next to the query, which is where anything finer is set.
What it is
A Tavily API client wrapping its search endpoints and its URL content extraction.
What you get
- A basic web search returning title, content and URL for every result.
- A context-aware search, tuned for retrieving contextually relevant results.
- A question-answering search, for direct questions that want a specific answer.
- Content extraction from a list of URLs, up to 20 of them in one call.
Requirements
TAVILY_API_KEY for a Tavily account.
Setup effort
One command plus a key — npx @mcptools/mcp-tavily, then supply credentials
USE CASES
01
Answering a factual question without opening ten tabs
The question-answering mode is aimed at exactly this: a direct question, a specific answer, no reading list.
02
Gathering background before writing about something
The plain search gives titles and URLs to work from; the context-aware one narrows the same ground to what is actually relevant.
03
Getting the text out of links you already collected
Hand over the URLs in a batch and take back processed content, without visiting each page.
4 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
search
Perform a basic web search using Tavily API. Returns search results including title, content and URL.
searchContext
Perform a context-aware web search using Tavily API. Optimized for retrieving contextually relevant results.
searchQNA
Perform a question-answering search using Tavily API. Best suited for direct questions that need specific answers.
BEFORE YOU INSTALL
What to know first
Tools you get
4 tools when we connected. Names are camelCase (searchContext, searchQNA). All four read: search is general, searchContext weights context, and searchQNA returns a direct answer.
Requires
- npx on PATH
- Node.js 16 or later
- TAVILY_API_KEY in your environment
Where it runs
The documented client runs the server on your machine.
CONTEXT COST
Light Standing cost
Rated by the size of the 4 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Server setup
TERMINAL
npx @mcptools/mcp-tavilyMCP Clients
Any MCP client
ANY MCP CLIENT · JSON
{
"mcpServers": {
"tavily": {
"command": "npx",
"args": [
"-y",
"@mcptools/mcp-tavily"
],
"env": {
"TAVILY_API_KEY": "<YOUR_TAVILY_API_KEY>"
}
}
}
}
