Skip to main content
CNAP’s documentation is automatically optimized for AI assistants, enabling you to get instant, accurate answers about CNAP directly from your favorite AI tools. The integration works through built-in MCP servers, contextual menus, and industry-standard llms.txt files with zero configuration required.

​
How AI Agents Work

CNAP provides multiple ways for AI tools to access documentation:
  • MCP server - Enables AI tools to search and interact with CNAP documentation in real-time. Available at https://docs.cnap.tech/mcp with no configuration needed
  • Contextual menu - Quick actions available on every documentation page for connecting to AI tools, copying content, and opening conversations
  • LLMs.txt files - Industry-standard files for efficient AI indexing, including a summary file and a complete documentation file

​
Key Features

  • Zero configuration - Automatic setup through contextual menu
  • Always current - Real-time data from latest documentation
  • Context-aware - AI understands your specific CNAP setup
  • Comprehensive - Access to all docs, APIs, and guides
  • Fast - Instant answers without manual searching

​
MCP Server

The Model Context Protocol (MCP) server enables AI tools to search and interact with CNAP documentation in real-time. The server is available at https://docs.cnap.tech/mcp and requires no configuration. AI tools can query the server to get up-to-date information about CNAP features, APIs, and documentation.

​
Contextual Menu

Every documentation page includes a contextual menu with quick actions for AI agents:
  • Copy page - Copy page content as Markdown for AI tools
  • View as Markdown - Open page as Markdown format
  • Open in AI tools - Create conversations with current page in ChatGPT, Claude, or Perplexity
  • Copy MCP server URL - Copy the MCP server URL to clipboard
  • Connect to editors - Auto-install MCP server in Cursor or VS Code

​
LLMs.txt Files

CNAP provides industry-standard llms.txt files for efficient AI indexing:
  • llms.txt - All pages with descriptions for quick reference
  • llms-full.txt - Complete documentation in one file for comprehensive access
These files help AI tools understand documentation structure and find relevant content quickly.

​
Supported AI Tools

CNAP’s AI agents work with various AI tools and editors:
  • Claude Desktop - Configure via settings file
  • Claude Connectors - Add as custom connector in Claude’s web interface
  • Claude Code - Add via command line interface
  • Cursor - Configure through MCP settings
  • VS Code - Requires MCP extension from marketplace
  • Other MCP-compatible tools - Configure via settings file using MCP server URL
Each tool can connect to the MCP server at https://docs.cnap.tech/mcp to access CNAP documentation.

​
Setting Up AI Agents

The easiest way to set up AI agents is through the contextual menu on any documentation page. Open the menu and select Connect to Cursor or Connect to VS Code for automatic installation. Your AI tool will then have access to CNAP documentation. If automatic setup isn’t available for your tool, you can configure the MCP server manually. The configuration varies by tool but typically involves adding the MCP server URL (https://docs.cnap.tech/mcp) to your tool’s configuration file.
Configuration File Location:
  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
  • Windows: %APPDATA%\Claude\claude_desktop_config.json
  • Linux: ~/.config/claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "cnap": {
      "command": "/Users/robin/.volta/bin/pnpm",
      "args": ["dlx", "mcp-remote", "https://docs.cnap.tech/mcp"],
      "env": {}
    }
  }
}
CNAP recommends pnpm for better caching and performance. The full path ensures compatibility across systems.
For Claude’s web interface, navigate to Claude Settings → Add custom connector, then configure:
  • Name: CNAP
  • URL: https://docs.cnap.tech/mcp
Use the attachments button (plus icon) → Select CNAP → Ask questions.
Add via command line:
Add MCP server
claude mcp add --transport http cnap https://docs.cnap.tech/mcp
Verify connection:
Verify connection
claude mcp list
Claude Code CLI is ideal for development workflows.
Command Palette:
  1. Press Cmd+Shift+P (Mac) or Ctrl+Shift+P (Windows/Linux)
  2. Search “Open MCP settings”
  3. Select Add custom MCP
mcp.json
{
  "mcpServers": {
    "cnap": {
      "url": "https://docs.cnap.tech/mcp"
    }
  }
}
Create .vscode/mcp.json:
.vscode/mcp.json
{
  "servers": {
    "cnap": {
      "type": "http",
      "url": "https://docs.cnap.tech/mcp"
    }
  }
}
Requires MCP extension from VS Code marketplace.
For other AI tools that support MCP, add the CNAP MCP server URL (https://docs.cnap.tech/mcp) to your tool’s configuration file. The exact format depends on your tool’s MCP implementation, but typically follows this pattern:
{
  "mcpServers": {
    "cnap": {
      "url": "https://docs.cnap.tech/mcp"
    }
  }
}
Configuration format may vary based on your tool’s MCP implementation. Refer to your tool’s documentation for the exact configuration structure.

​
Markdown Access

Access any documentation page as Markdown by appending .md to the URL. For example, https://docs.cnap.tech/ai-agents.md returns the page content in Markdown format, making it easy for AI tools to process and understand.

​
Example Questions

Once connected, you can ask your AI assistant about CNAP:
  • “How do I create a new product in CNAP?”
  • “What are the requirements for importing a cluster?”
  • “How does the payment system work?”
  • “What’s the difference between managed and imported clusters?”
  • “How do I set up pricing for my products?”
The MCP server provides real-time data for accurate, up-to-date answers.

​
Troubleshooting

  • Verify URL: https://docs.cnap.tech/mcp
  • Check internet connection
  • Ensure AI tool supports MCP protocol
  • Try contextual menu for automatic setup
  • Check JSON syntax in config files
  • Verify file permissions
  • Confirm correct file location
  • Restart your AI application
  • MCP server provides real-time data
  • Restart MCP connection
  • Clear cached responses in AI app
  • Verify correct MCP server URL