home / mcp / ravelry mcp server

Ravelry MCP Server

MCP for letting Claude search Ravelry

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "gpaul-mcp-mcp_ravelry": {
      "url": "http://localhost:3000",
      "headers": {
        "AUTH_PASS": "YOUR_RAVELRY_PASSWORD",
        "AUTH_USER": "YOUR_RAVELRY_USERNAME"
      }
    }
  }
}

You can run a dedicated MCP server that exposes Ravelry pattern search and detail tools for AI assistants. This server securely talks to the Ravelry API using your credentials and can be connected to MCP clients to browse and fetch knitting and crochet patterns.

How to use

You interact with the Ravelry MCP Server through an MCP client. Start the server locally or on a reachable host, then connect your client using the HTTP endpoint or the local stdio command. Once connected, you can search for knitting and crochet patterns, retrieve detailed information about specific patterns, and fetch details for multiple patterns at once. Use the client to issue queries like search by craft (knitting or crochet), filter by availability (free, paid), and request detailed pattern data for IDs you already have.

Typical actions you can perform include: searching for patterns with keywords, filtering by craft and price, getting full pattern details, and requesting multiple pattern details in a single call. The server handles authentication with your Ravelry credentials and returns structured results that your AI assistant can present to you.

How to install

Prerequisites you need before installing are: Node.js version 14 or higher and a working npm or yarn installation. You also need Ravelry API credentials (username and password). Make sure you have these ready before starting the setup.

Step 1: Prepare your environment
- Ensure Node.js is installed (v14 or higher)
- Install npm (or yarn) if not already available
- Have your Ravelry username and password ready for authentication

Step 2: Install dependencies and set up the project
- Create a working directory for the MCP server
- Install dependencies using npm
- Configure environment files for development and production

Step 3: Configure credentials - Create or copy environment files for development and production - Add your Ravelry credentials in both environments as AUTH_USER and AUTH_PASS

# Development setup
cp .env.example .env.development

# Production setup
cp .env.example .env.production

# Add credentials in both files
AUTH_USER=your_ravelry_username
AUTH_PASS=your_ravelry_password

Additional notes about configuration and operation

Development and production modes provide different startup flows. In development, you typically run the server with hot reloading for rapid iteration. In production, you build the server and then start it in a running state.

To connect with Claude Desktop or other MCP clients, you commonly supply either the server URL (for HTTP) or a local command to run the MCP server (for stdio), and you enable the MCP integration once the client can reach the server.

Configuration and connection options

The MCP server can be reached in two ways during setup. You can expose the server via HTTP so the client connects to a remote URL, or you can run the server locally as a stdio process and let the client launch and communicate with it.

Troubleshooting

If you cannot connect from your MCP client, verify that the server is running and accessible at the configured URL, check that the credentials are valid, and ensure the client is pointed to the correct endpoint. Confirm that the server has started successfully in development or production mode and that any required environment variables are set.

Available tools

search-patterns

Searches for knitting or crochet patterns based on query parameters such as keywords, craft, and availability with pagination support.

get-pattern-details

Retrieves detailed information for a single pattern by its ID, including metadata and crafting details.

get-multiple-pattern-details

Fetches detailed information for multiple patterns in a single request by providing an array of IDs.