MCP Server

The Hands-on AI Playbook is available as an MCP (Model Context Protocol) server, so you can get playbook-informed answers directly in your AI tools — no API keys or authentication required.

MCP server URL: https://mcp.handsonai.info/mcp

The Hands-on AI Playbook connector enabled in Claude. This works in any AI tool that supports MCP — see below for setup instructions for Claude, ChatGPT, Cursor, VS Code, and more.

What You Get

The MCP server gives your AI assistant access to:

• Search across all playbook pages by keyword
• Read the full content of any page
• Browse sections like Building Blocks, Framework, Use Cases, and more
• Deep dive into building blocks or framework steps with all sub-pages
• Q&A lookup with quick answers from the playbook’s question pages
• Setup guides for every tool in the Builder Tools

Connect from Claude

Claude supports remote MCP servers as custom connectors across all its products. Add the connector once, and it’s available in Claude chat, Claude Desktop, and Cowork.

Claude (claude.ai) and Claude Desktop

1. Go to Settings → Connectors
2. Scroll to the bottom and click Add custom connector
3. Enter the MCP server URL: https://mcp.handsonai.info/mcp
4. Click Add

To use it in a conversation, click the + button in the chat input area, select Connectors, and toggle the Hands-on AI Playbook connector on. The first time Claude uses the connector, you’ll see a permission prompt — click Always allow so it can search the playbook without asking each time.

You can manage permissions for each tool individually in Settings → Connectors → Hands-on AI Playbook. From there, set each tool to always allow, ask every time, or deny — giving you full control over what Claude can access.

Works on Pro, Max, Team, and Enterprise plans

Custom connectors are available on all paid Claude plans. Team and Enterprise owners configure connectors in Organization settings → Connectors, then members enable them individually in their personal settings.

Claude Cowork

Once you’ve added the connector using the steps above, it’s automatically available in Cowork. When Cowork runs a task, it can use the playbook connector to search for relevant content, look up setup guides, or reference building block definitions.

Claude Code

Run this command in your terminal:

claude mcp add handsonai-playbook --transport http https://mcp.handsonai.info/mcp

That’s it. Claude Code will now have access to all playbook tools in every conversation.

Connect from ChatGPT

ChatGPT supports remote MCP servers as apps. Available on Plus, Pro, Team, Enterprise, and Edu plans.

Enable Developer Mode (one-time setup):

1. Go to Settings → Apps
2. Under Advanced settings, toggle Developer Mode on

Add the app:

1. In Settings → Apps, click Create app
2. Fill in:
   • Name: Hands-on AI Playbook
   • URL: https://mcp.handsonai.info/mcp
3. Check the box confirming you trust this application
4. Click Create

Use it in a conversation:

1. Start a new chat
2. Click + → More → Developer Mode
3. Select the Hands-on AI Playbook app to enable it

Enable per conversation

The app must be explicitly added to each new chat session via Developer Mode. It works in both regular chat and Deep Research mode.

Connect from Cursor

Cursor supports MCP servers via the Command Palette or the Settings UI.

Option 1 — Command Palette:

Open the Command Palette (++cmd+shift+p++ on Mac, ++ctrl+shift+p++ on Windows/Linux) and search for View: Open MCP Settings. Cursor will open a mcp.json file — paste the configuration shown below, then save.

Option 2 — Settings UI:

1. Go to Settings → MCP
2. Click + Add new global MCP server
3. Cursor will open a mcp.json file — paste this configuration:

{
  "mcpServers": {
    "handsonai-playbook": {
      "url": "https://mcp.handsonai.info/mcp"
    }
  }
}

4. Save the file — Cursor will detect it and show the server in the MCP panel

You should see a green indicator next to “handsonai-playbook” in the MCP settings panel. The playbook tools will be available in Cursor’s Agent mode.

Connect from VS Code

VS Code (with GitHub Copilot) supports MCP servers via workspace or user settings.

Option 1 — User settings (available in all projects):

Open the Command Palette (++cmd+shift+p++ on Mac, ++ctrl+shift+p++ on Windows/Linux) and search for MCP: Add Server. Select HTTP as the transport type, enter https://mcp.handsonai.info/mcp as the URL, and name it handsonai-playbook.

Option 2 — Manual JSON configuration:

Add the following to your VS Code settings.json (user or workspace):

{
  "mcp": {
    "servers": {
      "handsonai-playbook": {
        "type": "http",
        "url": "https://mcp.handsonai.info/mcp"
      }
    }
  }
}

The playbook tools will be available in GitHub Copilot’s Agent mode (the @ chat mode).

Requires GitHub Copilot

MCP support in VS Code requires GitHub Copilot (Free, Pro, or Business). Make sure you have the GitHub Copilot extension installed and are signed in.

Connect from Other MCP Clients

Any MCP client that supports Streamable HTTP transport can connect:

• URL: https://mcp.handsonai.info/mcp
• Transport: Streamable HTTP (POST with JSON-RPC 2.0)
• Authentication: None required (public, read-only)

Available Tools

Tool	Description
search_playbook	Search pages by keyword, optionally filtered to a section
get_page	Get the full content of a specific page by path
list_section	List all pages in a section (building-blocks, framework, use-cases, etc.)
get_building_block	Get a building block and all its sub-pages (prompts, agents, mcp, etc.)
get_framework_step	Get a framework step and all its sub-pages (discover, deconstruct, build)
list_questions	List all Q&A pages with question text and short answers
get_setup_guide	Get a Builder Tools setup guide (terminal, editor, git, claude-code, etc.)

Example Prompts

Once connected, try asking your AI assistant:

• “What are the agentic building blocks?”
• “How do I set up Claude Code?”
• “Search the playbook for prompt engineering techniques”
• “What’s the Business-First AI Framework?”
• “Show me the automation use case”

The AI will use the playbook MCP tools to pull relevant content and answer with playbook knowledge.

Analytics & Privacy

No personally identifiable information is collected

This MCP server does not collect personally identifiable information. There are no user accounts, no IP addresses, no cookies, no tracking identifiers, and no conversation content stored — ever. We cannot identify who you are or what you asked your AI assistant.

What is logged: Each MCP tool call records the tool name, search keywords or page path requested, a timestamp, approximate country (from Cloudflare’s network), client type (User-Agent header), and response time. Only tool calls and resource reads are logged — protocol handshake messages are not.

Why: We use this aggregate data to understand which topics people search for most and where content gaps exist, so we can create new pages and improve existing ones. The goal is to make the playbook more useful for everyone.

Retention: Analytics data is retained for 12 months, then deleted.

Storage: Data is stored in a private database accessible only to the playbook maintainer. It is never shared, sold, or used for advertising.