Community MCP server for Tavily, the search engine built specifically for AI agents. Returns structured search results optimized for LLM consumption, including AI-generated summaries, extracted answers, and source citations. Supports depth control and content filtering. Superior to general web search for agent use cases where concise, accurate information retrieval is critical.
npx -y tavily-mcp
{
"mcpServers": {
"tavily": {
"env": {
"TAVILY_API_KEY": "your-api-key"
},
"args": [
"-y",
"tavily-mcp"
],
"command": "npx"
}
}
}Search web with Tavily AI
{
"type": "object",
"required": [
"query"
],
"properties": {
"query": {
"type": "string"
},
"max_results": {
"type": "integer"
},
"search_depth": {
"type": "string"
}
}
}// Input
{
"query": "MCP protocol specification",
"max_results": 5,
"search_depth": "advanced"
}// Output
{
"results": [
{
"url": "https://spec.modelcontextprotocol.io",
"title": "MCP Specification",
"content": "The Model Context Protocol..."
}
]
}Community MCP server for Tavily, the search engine built specifically for AI agents. Returns structured search results optimized for LLM consumption, including AI-generated summaries, extracted answers, and source citations. Supports depth control and content filtering. Superior to general web search for agent use cases where concise, accurate information retrieval is critical.
Tavily provides 1 tools including search.
Tavily offers a free tier with 200 calls per month. For higher usage, paid plans start at €0/month.
You can install Tavily using the following command: npx -y tavily-mcp. After installation, add the provided config snippet to your Claude Desktop or Cursor configuration.
Tavily is listed under the Data & Analytics category in the AgentForge MCP registry.
Tavily has a current uptime of 99.85% with an average response time of 350ms.
To connect Tavily, click the "Connect Agent" button on this page to get the configuration snippet. Add it to your MCP client (Claude Desktop, Cursor, or any MCP-compatible tool). Your AI agent will then have access to all of Tavily's tools via the Model Context Protocol.