Get the FREE Ultimate OpenClaw Setup Guide →

multi-search

Multi search MCP server

Installation
Run this command in your terminal to add the MCP server to Claude Code.
Run in terminal:
Command
claude mcp add --transport stdio guptabhishek-multi-search-mcp node build/index.js \
  --env PORT="3001" \
  --env GOOGLE_CX="YOUR_GOOGLE_CUSTOM_SEARCH_ENGINE_ID" \
  --env BRAVE_API_KEY="YOUR_BRAVE_API_KEY" \
  --env GOOGLE_API_KEY="YOUR_GOOGLE_API_KEY" \
  --env TAVILY_API_KEY="YOUR_TAVILY_API_KEY" \
  --env SEARCH_STRATEGY="priority" \
  --env SEARCH_PROVIDERS="tavily,google,duckduckgo,brave"

How to use

multi-search is a unified MCP tool that aggregates results from several web search providers (Tavily, Google Custom Search, DuckDuckGo, and Brave Search) and returns a standardized set of results. By default it uses a priority-based fallback between providers, but you can also specify a particular provider directly for a given query. The server exposes a single MCP tool named search, which accepts a query and optional parameters to control which provider to use and how many results to return. The output is a JSON object containing an array of content items with title, link, snippet, and provider information, making it easy to feed into downstream automations or AI workflows.

How to install

Prerequisites:

  • Node.js (LTS) and npm installed on your system
  • Access to the MCP client to configure environment variables

Install and run locally:

  1. Clone or download the repository for the multi-search MCP server.
  2. Install dependencies:
npm install
  1. Build the server (if the project uses a build step):
npm run build
  1. Start the server:
npm start

Configure using MCP client (example):

  1. Create or edit your MCP client configuration to spawn the server as a node process. Example (using build/index.js):
{
  "servers": {
    "search": {
      "autoApprove": [],
      "disabled": false,
      "timeout": 120,
      "command": "node",
      "args": ["build/index.js"],
      "env": {
        "TAVILY_API_KEY": "YOUR_TAVILY_API_KEY",
        "GOOGLE_API_KEY": "YOUR_GOOGLE_API_KEY",
        "GOOGLE_CX": "YOUR_GOOGLE_CUSTOM_SEARCH_ENGINE_ID",
        "BRAVE_API_KEY": "YOUR_BRAVE_API_KEY",
        "SEARCH_PROVIDERS": "tavily,google,duckduckgo,brave",
        "SEARCH_STRATEGY": "priority",
        "PORT": "3001"
      },
      "transportType": "stdio"
    }
  }
}
  1. Start your MCP client to launch the server through your configured command. When running locally, you’ll typically access the MCP tool named search via your client.

Additional notes

Environment variables allow you to enable/disable providers and control behavior. Important keys include TAVILY_API_KEY, GOOGLE_API_KEY, GOOGLE_CX, and BRAVE_API_KEY (the Brave key must be supplied via MCP env, not a local .env file). If a provider key is missing, that provider will be skipped. The default provider priority is Tavily, Google, DuckDuckGo, Brave, but you can customize via PROVIDER_LIST or the SEARCH_PROVIDERS setting in your MCP client configuration. The output format standardizes results so downstream tools can rely on consistent fields (title, link, snippet, provider). If you encounter rate limits, you can adjust the number of results or switch the strategy to random or prioritized fallback.

Related MCP Servers

Sponsor this space

Reach thousands of developers