Getting Started with Dwizi

Welcome! Dwizi turns your code into AI tools that work with Claude, ChatGPT, Cursor, and other AI systems. No infrastructure setup required.

What You'll Build

By the end of this guide, you'll have created an AI tool that can greet people. Your AI will be able to say things like:

"Hello! Can you greet Sarah for me?" → AI calls your tool → "Hello, Sarah!"

Step 1: Sign Up and Create Your First Tool

1. Go to the Dwizi Console: Visit app.dwizi.com and create an account
2. Click "Create Tool": You'll see a simple form to build your tool
3. Fill in the basics:
   • Name: "Greeting Tool"
   • Description: "A friendly greeting tool"
   • Language: Choose JavaScript (simpler) or TypeScript (more structured)

Step 2: Write Your First Function

Copy this code into the code editor:

export default async function greet(input) {
  const name = input?.name || "friend";
  return {
    message: `Hello, ${name}! Welcome to Dwizi.`
  };
}

What this does: Takes a name and returns a greeting message.

Step 3: Define What Your Tool Accepts

Dwizi needs to know what information your tool expects. In the "Input Schema" section, add:

{
  "type": "object",
  "properties": {
    "name": {
      "type": "string",
      "description": "The name of the person to greet"
    }
  }
}

This tells AI systems that your tool accepts a name parameter.

Step 4: Test Your Tool

1. Save your tool: Click "Create Tool"
2. Test it: Use the built-in tester in the console
3. Try it out: Put in a name like "Alice" and see the greeting

Your tool is now live and secure!

Step 5: Connect to AI (The Magic Part)

For Claude Desktop:

1. Open Claude Desktop on your computer
2. Go to Settings → Developer → Edit Config
3. Add your Dwizi MCP server:

{
  "mcpServers": {
    "dwizi": {
      "command": "node",
      "args": ["/path/to/dwizi-mcp-client"],
      "env": {
        "DWIZI_TOKEN": "your_token_here"
      }
    }
  }
}

For Cursor:

1. Open Cursor settings
2. Find MCP configuration
3. Add Dwizi server URL

For Other AI Systems:

Most AI systems that support MCP will have a configuration section. Just add your Dwizi MCP server URL.

Step 6: Use Your Tool with AI

Now you can ask your AI:

• "Greet Sarah"
• "Say hello to the team"
• "Create a welcome message for new users"

Your AI will automatically call your Dwizi tool and get the greeting!

What Just Happened?

You created a secure, isolated tool that:

• Runs in a container (like a secure box)
• Has resource limits (won't crash your system)
• Is accessible via URL (works from anywhere)
• Integrates with AI systems (MCP-compatible)
• Is ready for production (no more setup needed)

Next Steps

Build More Tools

Now that you know how, try these:

Weather Tool: Get current weather for any location Calculator: Perform mathematical operations Data Analyzer: Process spreadsheets or CSV files Email Sender: Send emails through your AI

Advanced Features

• Streaming: See results in real-time
• Private Tools: Require authentication
• Tool Groups: Combine multiple tools
• Organizations: Share tools with your team

Real-World Examples

Check out our use cases for inspiration:

• Business dashboards
• Content creation
• Data processing
• Developer tools

Need Help?

• Documentation: Explore our guides for detailed instructions
• Community: Join discussions about building AI tools
• Support: Contact us if you get stuck

Why Dwizi?

Instead of managing servers, containers, and security, you just write functions. Dwizi handles the infrastructure so you can focus on building amazing AI capabilities.

Welcome to the future of AI tooling!