home / mcp / mcp search server
Multi search MCP server
Configuration
View docs{
"mcpServers": {
"guptabhishek-multi-search-mcp": {
"command": "npx",
"args": [
"-y",
"mcp-search-server"
],
"env": {
"GOOGLE_CX": "YOUR_GOOGLE_CUSTOM_SEARCH_ENGINE_ID",
"BRAVE_API_KEY": "YOUR_BRAVE_SEARCH_API_KEY",
"GOOGLE_API_KEY": "YOUR_GOOGLE_API_KEY",
"TAVILY_API_KEY": "YOUR_TAVILY_API_KEY"
}
}
}
}You can run a unified MCP interface for multi-provider web search, letting you leverage free quotas from several search services in a single tool. This server standardizes results from multiple providers and lets you choose a provider directly or fall back automatically to maximize your available search capacity for automation and AI workflows.
You interact with a single MCP tool named search to perform web searches across multiple providers. You can either specify a particular provider to query or rely on a priority-based fallback to maximize available results. Use this to automate search-heavy workflows, run recurring checks, or feed search results into other AI tools.
Prerequisites: you need Node.js and npm installed on your machine.
Install or run the MCP server locally by following these steps.
Prepare the server configuration and start the server with the following commands.
The server can be configured by providing API keys for the supported providers and optional preferences for provider order and behavior.
If you encounter issues starting the server, verify that all required API keys are supplied and that the chosen port is not in use. Check that the server reports the correct tool name and that the API keys are recognized on startup.
Performs a web search using multiple providers (Tavily, Google, DuckDuckGo, Brave) with optional direct provider selection or priority-based fallback. Returns a structured list of results with provider attribution.