tavily_search
Search the web for current information on any topic. Use for news, facts, or data beyond your knowledge cutoff. Returns snippets and source URLs.
Web search built for agents, with domain filtering, plus page extraction, site crawling and mapping, and multi-source research.
LABSCO SUMMARY
Most agent searches fail not because the answer is missing but because the top results are content marketing. Pinning a search to the domains you trust — a vendor's docs, a regulator, a handful of publications — changes what comes back more than any prompt tuning will. One caution: passing the key as a URL parameter puts a credential into config files and shell history; prefer the header or OAuth where your client supports it.
Tavily's own hosted endpoint for agent web research. Their search is built for retrieval by models rather than for human browsing, which shows in the filtering.
Paste a URL, then authorize — add the endpoint to your client, then approve the OAuth consent screen
USE CASES
Restrict to specific domains rather than filtering the noise afterwards.
Extraction that returns the content, not the chrome.
Map a site's URLs, then crawl the sections you need with plain-language instructions.
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
tavily_searchSearch the web for current information on any topic. Use for news, facts, or data beyond your knowledge cutoff. Returns snippets and source URLs.
tavily_extractExtract content from URLs. Returns raw page content in markdown or text format.
tavily_crawlCrawl a website starting from a URL. Extracts content from pages with configurable depth and breadth.
tavily_mapMap a website's structure. Returns a list of URLs found starting from the base URL.
tavily_researchPerform comprehensive research on a given topic or question. Use this tool when you need to gather information from multiple sources, including web pages, documents, and other resources, to answer a question or complete a task. Returns a detailed response based on the research findings. Rate limit: 20 requests per minute.
BEFORE YOU INSTALL
5 tools when we connected.
CONTEXT COST
Rated by the size of the 5 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Add the endpoint to your client, then approve the OAuth consent screen.
https://mcp.tavily.com/mcp/claude mcp add --transport http tavily https://mcp.tavily.com/mcp/?tavilyApiKey=<YOUR_TAVILY_API_KEY>claude mcp add --transport http tavily https://mcp.tavily.com/mcpclaude mcp add --transport http --scope user tavily https://mcp.tavily.com/mcp/?tavilyApiKey=<YOUR_TAVILY_API_KEY>{
"mcpServers": {
"tavily-remote-mcp": {
"command": "npx -y mcp-remote https://mcp.tavily.com/mcp/?tavilyApiKey=<YOUR_TAVILY_API_KEY>",
"env": {}
}
}
}{
"mcpServers": {
"tavily-remote-mcp": {
"command": "npx mcp-remote https://mcp.tavily.com/mcp",
"env": {}
}
}
}{
"mcpServers": {
"tavily-mcp": {
"command": "npx",
"args": [
"-y",
"tavily-mcp@latest"
],
"env": {
"TAVILY_API_KEY": "<YOUR_TAVILY_API_KEY>"
}
}
}
}Connect to the Tavily Remote MCP with the keyless header.
claude mcp add tavily-remote-mcp --transport http https://mcp.tavily.com/mcp/ \
--header "X-Tavily-Access-Mode: keyless"The header is required — without it the server asks your client to sign in. Clients that only accept a URL can't send it; use an API key with those.