CRC Documentation MCP Server

An MCP (Model Context Protocol) server that provides intelligent access to CRC/OpenShift Local documentation. This server fetches, caches, and searches through official CRC documentation to answer questions about CRC or OpenShift Local.

Features

• Multi-source Documentation Access: Fetches content from multiple CRC documentation sources

• Intelligent Search: Finds relevant sections based on query relevance scoring

• Caching System: Reduces API calls by caching fetched documentation

• Clean Text Extraction: Extracts readable content from HTML pages

• MCP Integration: Works seamlessly with MCP-compatible clients

Related MCP server: Atlas

Documentation Sources

The server accesses the following CRC documentation sources:

• CRC Docs: https://crc.dev/docs - Official documentation

• CRC Blog: https://crc.dev/blog - Blog posts and announcements

• CRC Engineering: https://crc.dev/engineering-docs - Engineering documentation

Installation

Prerequisites

• Python 3.13 or higher

• uv (recommended) or pip

Install with uv (recommended)

# Clone the repository
git clone <repository-url>
cd crc-documentation

# Install dependencies
uv sync

Usage

Running the Server

Start the MCP server:

# With uv
uv run server.py

# With Python directly
python server.py

The server runs on stdio transport and is designed to be used with MCP-compatible clients.

Using with Cursor/VSCode

Cursor Configuration

To use this MCP server with Cursor, add the following configuration to your Cursor settings:

1. Open Cursor Settings (Cmd/Ctrl + ,)

2. Search for "MCP" or go to Extensions → MCP

3. Add a new MCP server configuration:

{
  "mcpServers": {
    "crcdocs": {
      "command": "/path/to/your/uv/binary",
      "args": [
        "--directory",
        "/path/to/your/crc-documentation",
        "run",
        "server.py"
      ]
    }
  }
}

Deployed Server

Cursor Configuration

{
  "mcpServers": {
    "crcdocs": {
      "url": "https://crc-doc-mcp.onrender.com/mcp/"
    }
  }
}

Testing the Integration

Once configured, you can test the integration by:

1. Restart Cursor/VSCode

2. Open the MCP panel or use the command palette

3. Try asking questions like:

   • "How do I install CRC on Linux?"

   • "What are the system requirements for CRC?"

   • "How do I troubleshoot CRC startup issues?"

The server will search through CRC documentation and provide relevant answers.

Available Tools

crc_doc_query

Query CRC documentation to get answers about CodeReady Containers or OpenShift Local.

Parameters:

• query (required): Your question about CRC or OpenShift Local

• sources (optional): List of specific sources to search (defaults to all sources)

Example:

# Query all sources
result = await crc_doc_query("How do I install CRC on Linux?")

# Query specific sources
result = await crc_doc_query(
    "What are the system requirements?", 
    sources=["crc", "crc-blog"]
)

clear_cache

Clear the documentation cache to fetch fresh content on the next query.

Example:

result = await clear_cache()

Configuration

The server is configured through the following constants in server.py:

DOC_SOURCES = {
    "crc": "https://crc.dev/docs",
    "crc-blog": "https://crc.dev/blog", 
    "crc-dev": "https://crc.dev/engineering-docs"
}

How It Works

1. Content Fetching: The server fetches HTML content from CRC documentation sources

2. Text Extraction: Uses BeautifulSoup to extract clean, readable text from HTML

3. Relevance Scoring: Analyzes content to find sections most relevant to your query

4. Caching: Stores fetched content to improve performance on subsequent queries

5. Response Formatting: Returns structured results with source attribution

Contributing

1. Fork the repository

2. Create a feature branch

3. Make your changes

4. Add tests if applicable

5. Submit a pull request

Logging

The server uses Python's built-in logging module. Logs are set to INFO level by default and include:

• Documentation fetching operations

• Error handling for failed requests

• Cache operations

Support

For issues with this MCP server, please open an issue in the repository.