Websearch MCP Server

Real-time web search capabilities for AI models using a dedicated crawler service.

Data last scanned yesterday · Reviewed last month

151 tool★ 38JavaScript

Overview

WebSearch-MCP allows AI assistants to fetch live information from the internet by integrating with a secondary Crawler API service. It communicates through the Model Context Protocol (MCP) over stdio, providing real-time data retrieval for models like Claude Desktop, Cursor, and Cline.

Our verdict

With a security score of 15/100, this tool is categorized as high risk. The repository has not received updates for over 454 days and lacks a declared open-source license, creating significant uncertainty regarding maintenance and legal usage terms. Due to these factors, it is recommended only for sandboxed development environments where stale dependencies and lack of oversight do not pose a threat to production data.

Supports stdio transport for MCP-compliant AI clients.
Configurable maximum search result limit.
Small dependency footprint with only 3 direct packages.

Last commit occurred over 454 days ago.
No license detected, creating legal ambiguity.
Community-maintained without official vendor oversight.

Setup

Claude Desktop

{
  "mcpServers": {
    "websearch-mcp": {
      "args": [
        "-y",
        "websearch-mcp"
      ],
      "command": "npx"
    }
  }
}

Paste into ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows), then restart Claude Desktop. Some servers require API keys or tokens as environment variables — check the project's README and add an "env" object if needed.

No read-only mode — running this grants full read/write access.
No built-in authentication — anyone with access to this config can invoke every tool it exposes.

Tools

Tool	Description	Risk
web_search	Performs a web search to retrieve up-to-date information about a topic using the configured Crawler API.	low

Compatibility

Client	Local	Docker	Remote	Read-only
ChatGPT
Claude Desktop
Cursor
VS Code
Windsurf

Frequently asked questions

›What infrastructure is required to run the WebSearch MCP server?

You must run a separate crawler service alongside the MCP server, which is typically set up using Docker Compose with the provided images for the crawler and FlareSolverr.

›Does the server require any specific configuration to connect to the crawler API?

The server connects to the crawler service via the API_URL environment variable, which defaults to http://localhost:3001 if not specified.

›Can I control the number of search results returned by the AI?

Yes, you can set the default maximum number of search results using the MAX_SEARCH_RESULT environment variable, which defaults to 5.

›How can I troubleshoot if the search functionality is not working?

You should first verify the crawler service health by running a curl request against its health endpoint and ensure that the FlareSolverr service is properly connected and reporting a status of true.

›Are there specific hardware considerations for running the crawler service?

If you are using a Mac with Apple Silicon, you should use the platform-specific docker-compose configuration provided in the documentation to ensure the crawler and FlareSolverr services run correctly.