MCP Tavily

MCP Tavily is an MCP server that connects large language model clients to the Tavily search engine API to provide search and web extraction capabilities. Designed for researchers, developers, and AI agent builders, it equips LLMs with tools to retrieve current information across the web, extract raw article content from target URLs, and filter results by domain, date, or category. By bridging Tavily with the Model Context Protocol, the server allows AI models to query news, finance, or general topics in real time, bypassing training knowledge cutoffs. Users can run standard queries, extract full-text content with or without images, and perform question-answering search routines that provide synthesized answers directly from Tavily. The server supports deep configuration options, such as basic or advanced search and extraction depths, time-range constraints, and token limits, helping agents find relevant, structured web data for automated research and synthesis tasks without requiring custom web scrapers.

Category: Browser & Web Automation

Tags: research, tavily, web scraping, web search

Visit MCP Tavily

How to install and configure MCP Tavily

1. Obtain an API key from Tavily. 2. To install automatically for Claude Desktop via Smithery, run: bash npx -y @smithery/cli install @kshern/mcp-tavily --client claude 3. Alternatively, configure the server manually in your MCP client configuration file: json { "mcpServers": { "tavily": { "command": "npx", "args": ["-y", "@mcptools/mcp-tavily"], "env": { "TAVILY_API_KEY": "your-api-key" } } } } 4. Replace your-api-key with your valid Tavily API key and restart the client.

What you can do with MCP Tavily

• Searching recent news and financial updates by specifying topic filters and time ranges like day, week, or month. * Extracting clean article text and image references directly from multiple web URLs without managing custom scraping infrastructure. * Retrieving direct answers to factual queries using Tavily Q&A search to provide cited information to users. * Restricting search queries to specific allowed domains or excluding untrusted sources to ensure high-quality research outputs. * Gathering comprehensive contextual research using advanced search depth to synthesize technical or industry briefings.

Key facts

• https://github.com/kshern/mcp-tavily
• Browser & Web Automation, Web Search & Research
• research, tavily, web scraping, web search

Part of MCP Servers

What is MCP Tavily?

MCP Tavily is a Model Context Protocol server that integrates the Tavily API into MCP-compatible clients. It provides automated tools for AI assistants to execute advanced web searches, run domain-filtered context queries, generate Q&A responses, and extract text and images from web pages.

How do I install MCP Tavily?

You can install the server automatically for Claude Desktop using Smithery by running npx -y @smithery/cli install @kshern/mcp-tavily --client claude. Alternatively, configure your client JSON settings to run npx -y @mcptools/mcp-tavily and provide your Tavily API key through the TAVILY_API_KEY environment variable.

What search tools does MCP Tavily provide?

The server provides four main tools: search for basic and advanced keyword queries, searchContext for topic and time-bounded inquiries, searchQNA for direct question-answering with synthesized answers, and extract for retrieving raw text and images from a list of web page URLs.

Which MCP clients work with MCP Tavily?

MCP Tavily works with any client supporting the Model Context Protocol, including Claude Desktop, Cursor, and development tools like MCP Inspector. Clients must support standard stdio MCP servers configured via command-line execution and environment variables.