home / mcp / crew research mcp server
Executes multi-agent deep research tasks by coordinating LinkUp searches, Gemini reasoning, and CrewAI orchestration through an MCP stdio endpoint.
Configuration
View docs{
"mcpServers": {
"ragha02-deep_rsrch_gemini": {
"command": "uv",
"args": [
"--directory",
"./Multi-Agent-deep-researcher-mcp-windows-linux",
"run",
"server.py"
],
"env": {
"GEMINI_API_KEY": "your_gEMINI_API_KEY",
"LINKUP_API_KEY": "your_linkup_api_key_here"
}
}
}
}You can run an MCP-powered multi-agent deep researcher that orchestrates agents with CrewAI to perform deep web searches via LinkUp, using Gemini for reasoning and a Streamlit UI for interaction. This server exposes multi-agent search capabilities that you can connect to from an MCP client to run complex research tasks end-to-end.
You will connect to the Crew Research MCP server from your MCP client. The server runs locally as a stdio MCP endpoint, launched with the runtime tool and a specific working directory. It orchestrates agents to perform web searches and reason over the results using Gemini, with results surfaced through a lightweight UI powered by Streamlit.
Prerequisites you need before starting: - Node.js or Python runtime for client interactions is not required for the server itself, but you will need an MCP runtime to launch stdio endpoints (the following commands assume the MCP runtime is available in your environment). - Access keys for the services the server uses: a LinkUp API key and a Gemini API key.
Configuration is provided to run the MCP server as a stdio endpoint named crew_research. The server is started by invoking the MCP runtime with a specific directory and entry script. You must supply your LinkUp and Gemini API keys in the environment before starting the server.