home / mcp / deep search mcp server

Deep Search MCP Server

Provides multi-provider web search via a single MCP server with configurable API keys.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "just-every-mcp-deep-search": {
      "command": "npx",
      "args": [
        "-y",
        "@just-every/mcp-deep-search"
      ],
      "env": {
        "ENV_FILE": "YOUR_PATH/.llm.env"
      }
    }
  }
}

MCP Deep Search is a server that lets you perform web searches across multiple providers through a single MCP interface. It lets you query engines like Brave, Google, OpenAI, and more, and can run locally or via an MCP client to streamline automated research workflows.

How to use

You connect to the MCP server using an MCP client, then run searches through the deep_search tool to pull results from a chosen provider or a comprehensive multi-source workflow. Use it to fetch web results for specific queries, or to perform deeper research when you want AI-assisted insights across several search engines.

How to install

Prerequisites: you need Node.js installed to run the MCP server locally. You also need an MCP client ( Claude or another MCP client) that can consume MCP servers and pass environment variables.

Additional sections

Configuration and runtime for the MCP server are shown below. You can run the server as a local stdio service using npx, with an environment file to provide API keys and credentials.

{
  "mcpServers": {
    "deep_search": {
      "command": "npx",
      "args": ["-y", "@just-every/mcp-deep-search"],
      "env": {
        "ENV_FILE": "/path/to/.llm.env"
      }
    }
  }
}

Available tools

deep_search

Performs web searches using a specific provider with options for query, provider, maxResults, and includeAnswer.

comprehensive_research

Executes AI-assisted, multi-engine research to compile a broad set of results and insights.