What is Openrouter Search Server

openrouter-search-server is an MCP server designed to provide web search capabilities by utilizing the OpenRouter API.

Use cases

Use cases include integrating web search capabilities into applications, enhancing search functionalities in development environments, and providing AI-driven search results for user queries.

How to use

To use openrouter-search-server, clone the repository, install dependencies, build the TypeScript code, and configure your MCP client with the necessary API key and settings.

Key features

Key features include web search via OpenRouter, use of the specific model ‘google/gemini-2.5-pro-exp-03-25:free:online’, raw text output from the AI model, and MCP integration with a ‘web_search’ tool.

Where to use

openrouter-search-server can be used in various fields that require web search functionality, such as software development, research, and data analysis.

Attention

Archived repo. For future developments see: https://github.com/joaomj/openrouter-websearch-mcp

OpenRouter Search MCP Server

This is a Model Context Protocol (MCP) server designed to provide web search capabilities by leveraging the OpenRouter API.

Features

• Web Search via OpenRouter: Utilizes the OpenRouter API to perform web searches.
• Specific Model: Uses the google/gemini-2.5-pro-preview-03-25 model.
• Raw Text Output: The server returns the direct, unparsed text response generated by the AI model.
• MCP Integration: Exposes a simple web_search tool compatible with MCP clients.

Installation

1. Clone: Clone this repository to your local machine.

   git clone https://github.com/joaomj/openrouter-search-server.git
   cd openrouter-search-server
   

2. Install Dependencies:

   npm install
   

3. Build: Compile the TypeScript code.

   npm run build
   

Configuration

1. API Key: Ensure you have an API key from OpenRouter.

2. MCP Settings: Add the following configuration block to your MCP client’s settings file (e.g., ~/.vscode-server/data/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json for VS Code Remote/Codespaces or ~/Library/Application Support/Claude/claude_desktop_config.json for Claude Desktop App):

   • Replace /home/<user>/Cline/MCP/openrouter-search-server/build/index.js with the actual absolute path to the built index.js file on your system if you cloned it elsewhere.
   • Replace YOUR_OPENROUTER_API_KEY_HERE with your actual OpenRouter API key.

   Important: The server requires the OPENROUTER_API_KEY environment variable to be set. Never commit your API key directly into the code or the repository.

Usage

Once the server is configured and running (either manually via npm start in its directory or automatically via your MCP client), you can use the web_search tool.

Tool: web_search

Server Name: openrouter-search (or whatever key you used in the MCP settings)

Input:

• query (string, required): The search query you want to perform.

Output:

• A text string containing the raw response from the OpenRouter API’s google/gemini-2.5-pro-preview-03-25 model.

Example MCP Tool Call:

<use_mcp_tool>
<server_name>openrouter-search</server_name>
<tool_name>web_search</tool_name>
<arguments>
{
  \"query\": \"What is the latest version of the MCP SDK?\"
}
</arguments>
</use_mcp_tool>

License

MIT.