Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Simple MCP Demo

This example provides the quickest way to get started with MCP (Model Context Protocol) integration, demonstrating how to expose basic SCIM operations to AI agents with minimal setup. It's perfect for understanding MCP concepts and testing AI agent interactions.

What This Example Demonstrates

  • Minimal MCP Setup - Get AI agents working with SCIM in under 50 lines of code
  • Basic Tool Exposure - Essential user management operations as AI tools
  • Simple Protocol Integration - Standard I/O based MCP communication
  • Quick Testing Patterns - Immediate feedback and validation for AI interactions
  • Foundation Building - Starting point for more sophisticated AI integrations

Key Features Showcased

Streamlined Integration

See how ScimMcpServer can be set up with minimal configuration, focusing on core functionality rather than advanced features.

Essential Tool Set

The example exposes a focused set of AI tools covering the most common identity management operations:

  • User creation and retrieval
  • Basic user queries
  • Schema discovery for AI understanding

Zero-Configuration AI Support

Watch how AI agents can immediately start working with your SCIM server without complex setup, authentication, or protocol negotiation.

Interactive Testing

The demo is designed for immediate interaction, allowing you to test AI agent communication patterns and understand the request/response flow.

Concepts Explored

This example introduces fundamental MCP concepts:

Perfect For Getting Started

This example is ideal if you're:

  • New to MCP - Understanding AI agent integration concepts
  • Rapid Prototyping - Quick setup for testing AI workflows
  • Proof of Concept - Demonstrating AI-driven identity management
  • Learning Integration - Understanding how SCIM and AI agents work together

Tool Capabilities

The simple demo exposes core identity management tools:

User Management

  • Create User - Provision new accounts with basic validation
  • Get User - Retrieve user information by username or ID
  • List Users - Browse available user accounts

System Discovery

  • Server Info - Basic server capabilities and configuration
  • Schema Info - Available resource types for AI understanding

AI Interaction Flow

The example demonstrates a typical AI agent workflow:

  1. Tool Discovery - AI agent requests available tools
  2. Schema Understanding - Agent learns about user attributes and validation
  3. Operation Execution - Agent performs identity management tasks
  4. Result Processing - Agent receives structured responses for decision making

Running the Example

cargo run --example simple_mcp_demo --features mcp

The server starts in interactive mode, ready to receive MCP protocol messages and demonstrate AI agent communication patterns.

Testing with AI Agents

Once running, you can test with various AI systems:

Manual Protocol Testing

Send JSON-RPC messages directly to understand the protocol:

echo '{"jsonrpc":"2.0","method":"tools/list","id":1}' | cargo run --example simple_mcp_demo --features mcp

AI Agent Integration

Connect with MCP-compatible AI agents to see natural language identity management in action.

Key Differences from Full MCP Server

This simplified demo differs from the full MCP server example:

  • Reduced Tool Set - Focus on essential operations only
  • Minimal Configuration - Default settings for quick startup
  • No Multi-Tenancy - Single-tenant operation for simplicity
  • Basic Error Handling - Simple error responses without complex recovery

Extending the Demo

Natural extensions to explore:

  • Additional Tools - Add group management or advanced user operations
  • Authentication - Integrate with your authentication system
  • Multi-Tenancy - Add tenant context for enterprise scenarios
  • Custom Schemas - Extend with organization-specific attributes

Running the Example

The demo starts immediately and provides clear output showing:

  • Available tools and their schemas
  • Example AI agent interactions
  • Request/response patterns
  • Error handling demonstrations

Next Steps

After exploring the simple demo:

Source Code

View the complete implementation: examples/simple_mcp_demo.rs