RAG Web Browser

Provides web search and page fetch capabilities for MCP-enabled AI agents via the RAG Web Browser.

typescript

0

GitHub Stars

typescript

Language

7 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "mcp-mirror-apify_mcp-server-rag-web-browser": {
      "command": "npx",
      "args": [
        "/path/to/mcp-server-rag-web-browser/build/index.js"
      ],
      "env": {
        "APIFY-API-TOKEN": "your-apify-api-token"
      }
    }
  }
}

You can run the MCP server for the RAG Web Browser to empower AI agents with web search and page-fetch capabilities. It connects through the Model Context Protocol, enabling secure, controlled interactions between your MCP client and the local server to crawl the web and return content in Markdown for use in prompts and analyses.

How to use

You interact with the RAG Web Browser MCP server through a compatible MCP client. The server exposes a single core capability: the search tool. You can issue a search query to the server to perform a web search, have it crawl the top results, and return the cleaned content of those pages as Markdown. You can also request the content of a single URL, which the server will fetch and return in Markdown. Use these capabilities to gather up-to-date information from the web and feed it into your AI workflows.

How to install

Prerequisites you need before installing this MCP server are listed below. Ensure you have a supported operating system, Node.js, and the required API token for web crawling services.

Prerequisites include the following:

• MacOS or Windows • Node.js (v18 or higher) • Apify API Token (APIFY_API_TOKEN) • Claude Desktop (or another MCP client that supports MCP servers)

Install and configure the MCP server for use with your MCP client as follows.

Install Claude Desktop and prepare to recognize the MCP server. The configuration should reference the MCP server entry as shown in the example below.
Restart Claude Desktop after editing the configuration to ensure the new MCP server is detected. Look for the connection indicator to confirm the server is connected.

Configuration and start commands

The MCP server is configured as a local stdio server that runs via a command line entry. The following configuration shows how the MCP client should launch the server using npx and a path to the built index.js.

{
  "mcpServers": {
    "mcp-server-rag-web-browser": {
      "command": "npx",
      "args": [
        "/path/to/mcp-server-rag-web-browser/build/index.js"
      ],
      "env": {
        "APIFY-API-TOKEN": "your-apify-api-token"
      }
    }
  }
}

After you configure the server, you can start Claude Desktop and the MCP client will connect to the local server via stdio. The server will expose the search tool to perform web searches and fetch page content.

Additional notes

For development and troubleshooting, you can test the server locally with the built-in client example and debug utilities. You can run a local test client to verify the server exposes the available tools and to exercise the search capability.

Available tools

search

Query Google Search, scrape the top N URLs from the results, and returns their cleaned content as Markdown. This tool accepts a query string and an optional max_results number.