Skip to main content
This guide walks you through configuring pdf noodle’s MCP server for each supported AI assistant. Choose the setup that matches your environment.

Prerequisites

Before you begin, you’ll need:
1

pdf noodle Account

Create an account if you don’t have one
2

API Key

Get your API key from API Settings
3

AI Assistant

Have one of the supported AI assistants installed and ready

Claude Desktop

Claude Desktop supports both local and remote MCP servers.

Cursor

Cursor IDE works great with the local MCP server for secure, offline-capable PDF generation.

ChatGPT

ChatGPT only supports remote MCP servers through its Connected Apps feature.
1

Open ChatGPT Settings

Click your profile icon → Settings
2

Navigate to Connected Apps

Go to Connected Apps or MCP Servers section
3

Add New Server

Click “Add Server” or “Connect App”
4

Configure the Connection

  • Name: pdf noodle
  • URL: https://mcp.pdfnoodle.com/mcp?api_key=your-api-key-here
  • Authentication: None (API key is in the URL)
5

Save and Test

Save the configuration and try: “List my PDF templates”
ChatGPT’s MCP support may vary by subscription tier and region. Check OpenAI’s documentation for the latest availability.

Claude.ai (Web)

Use MCP with Claude directly in your browser.
1

Open Claude Settings

Click your profile → Settings
2

Go to Developer Settings

Navigate to DeveloperMCP Servers
3

Add pdf noodle

  • Server URL: https://mcp.pdfnoodle.com/mcp?api_key=your-api-key-here
  • Name: pdf noodle
4

Enable the Server

Toggle the server on and start a new conversation

Windsurf

Windsurf supports MCP through its configuration system. The local server is recommended for better security and offline access.

n8n

Integrate pdf noodle MCP into automated workflows using n8n. n8n connects to the remote MCP server via HTTP requests.
1

Create a New Workflow

Open n8n and create a new workflow or edit an existing one
2

Add an HTTP Request Node

Click the + button and search for “HTTP Request”
3

Configure the Request

Set up the HTTP Request node with these settings:
SettingValue
MethodPOST
URLhttps://mcp.pdfnoodle.com/mcp?api_key=your-api-key-here
AuthenticationNone
Content-Typeapplication/json
Authentication is set to None because the API key is passed directly in the URL query parameter.
4

Set the Request Body

Switch to the Body tab and select JSON. Use the JSON-RPC 2.0 format:
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "html_to_pdf",
    "arguments": {
      "html": "<h1>Hello from n8n!</h1>",
      "pdfParams": "{\"format\": \"A4\"}"
    }
  }
}
5

Execute and Test

Click Test step to run the request. You should receive a response with the PDF download URL.

Available MCP Methods for n8n

MethodDescriptionExample params.name
tools/callExecute a toolhtml_to_pdf, generate_pdf, list_templates
tools/listList available tools

Example: Generate PDF from Template

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "generate_pdf",
    "arguments": {
      "templateId": "your-template-id",
      "data": "{\"customerName\": \"John Doe\", \"invoiceNumber\": \"INV-001\"}"
    }
  }
}

Example: List All Templates

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "list_templates",
    "arguments": {}
  }
}
For production workflows, store your API key securely using n8n credentials or environment variables instead of hardcoding it in the URL.
You can chain multiple n8n nodes together to create complex workflows—for example, fetch data from a database, generate a PDF, and send it via email.

Manual Execution

Run the MCP server directly from the command line for testing or debugging.
# Set your API key
export PDFNOODLE_API_KEY=your-api-key-here

# Run the server
npx mcp-server-pdfnoodle
The server will start and listen for MCP protocol messages on stdin/stdout.

Verifying Your Setup

After configuration, verify everything works:
1

Start a New Conversation

Open your AI assistant and start a fresh conversation
2

Test the Connection

Ask: “List my pdf noodle templates”
3

Check the Response

You should see a list of templates or a message that no templates exist yet
If you encounter issues, check the Troubleshooting section below.

Troubleshooting

  • Verify your API key is correct
  • Check that the configuration file syntax is valid JSON
  • Ensure you’ve restarted the AI assistant after configuration changes
  • For local servers, verify Node.js is installed (node --version)
  • Double-check your API key at app.pdfnoodle.com/settings/api
  • Ensure there are no extra spaces in the API key
  • Verify your account is active and has available credits
  • Some AI assistants cache tool lists—try restarting completely
  • Check if MCP is enabled in your assistant’s settings
  • Verify the server name matches in your configuration
  • Check your account has sufficient credits
  • For HTML-to-PDF, ensure the HTML is valid
  • Review the PDF Best Practices for formatting guidelines

Security Considerations

API Key Protection

  • Never commit API keys to version control
  • Use environment variables when possible
  • Rotate keys periodically

Network Security

  • Remote URLs include your API key—use HTTPS only
  • Prefer local servers for sensitive environments
  • Review your assistant’s data handling policies

Next Steps