Skip to main content
Welcome to the Koog Framework Examples! This collection showcases real-world implementations of AI agents built with Koog, from simple calculators to production-ready applications.

​
Available Examples

Each example demonstrates different aspects of the Koog framework, with complete source code and step-by-step explanations.

​
Simple Agent Examples

Simple Agent Examples

A comprehensive collection of runnable examples demonstrating core Koog concepts:
  • Calculator agents with tool calling
  • Banking assistant with routing
  • Structured output and streaming
  • Memory and persistence patterns
  • OpenTelemetry integration
Difficulty: Beginner to Intermediate

​
Code Agent Tutorial

Building a Code Agent

A progressive, 5-step tutorial building a production-grade coding agent:
  • Step 1: Minimal agent with file operations
  • Step 2: Add shell command execution
  • Step 3: Integrate observability (Langfuse)
  • Step 4: Add specialized sub-agents
  • Step 5: Implement history compression
Difficulty: Intermediate

​
Trip Planning Agent

Trip Planning Agent

An advanced agent that plans trips by integrating multiple APIs:
  • Google Maps via MCP (Model Context Protocol)
  • Weather forecasts from Open Meteo
  • Interactive user conversations
  • Complex graph-based strategy
Difficulty: Advanced

​
ACP Agent Integration

ACP Agent for IDEs

Connect Koog agents to IntelliJ IDEA using the Agent Communication Protocol:
  • File operations and code navigation
  • IDE integration patterns
  • Bidirectional communication
  • Session management
Difficulty: Advanced

​
Spring Boot Integration

Spring Boot Integration

Production-ready Spring Boot application with Koog agents:
  • REST API for chat interactions
  • MCP server integration (GitHub)
  • S3 persistence for agent state
  • Configurable via YAML
Difficulty: Intermediate

​
Quick Start

All examples are available in the Koog GitHub repository: 
git clone https://github.com/koogio/koog.git
cd koog/examples

​
Prerequisites

Before running the examples, ensure you have:
  • Java 17+ installed
  • Kotlin 1.9+ (included with Gradle)
  • API Keys for LLM providers:
    • OpenAI API key (most examples)
    • Anthropic API key (optional)
    • Google AI API key (optional)

​
Setting Up API Keys

Set your API keys as environment variables: 
export OPENAI_API_KEY="your_openai_key"
export ANTHROPIC_API_KEY="your_anthropic_key"
export GOOGLE_AI_API_KEY="your_google_key"
Or create an env.properties file in your project root: 
OPENAI_API_KEY=your_openai_key
ANTHROPIC_API_KEY=your_anthropic_key
GOOGLE_AI_API_KEY=your_google_key

​
Example Difficulty Levels

Examples with minimal complexity, perfect for learning basic Koog concepts:
  • Simple calculator agents
  • Basic tool calling
  • Single-run strategies
Examples demonstrating more complex patterns:
  • Multi-agent routing
  • Event handling and observability
  • Structured output
  • State persistence
Production-ready implementations with advanced features:
  • Complex graph strategies
  • Multi-API integration
  • MCP protocol integration
  • History compression
  • Sub-agent orchestration

​
Learning Path

We recommend exploring examples in this order:
  1. Start with Simple Agent Examples to understand core concepts
  2. Follow the Code Agent Tutorial to build a complete agent step-by-step
  3. Explore Trip Planning for advanced multi-API integration
  4. Try Spring Boot Integration for production deployment
  5. Experiment with ACP Agent for IDE integration

​
Additional Resources

  • Interactive Notebooks: Jupyter notebooks for hands-on learning in the examples/notebooks/ directory
  • API Reference: Detailed API documentation at api.koog.ai
  • Community: Join discussions on GitHub

​
Next Steps

Simple Agent Examples

Start with basic examples

Code Agent Tutorial

Build a coding agent step-by-step

Build docs developers (and LLMs) love

Get started for freeTalk to us