Learn Strands Agent Framework - Hands-On Tutorial

Learn by Building: Master the Strands Agent Framework through 10 comprehensive, hands-on lessons covering everything from basics to production deployment.

🎯 What You'll Learn

Master the Strands Agent Framework through 10 progressive lessons teaching production-ready agent patterns used by leading AI companies.

Your Learning Journey (25-30 hours)

%%{init: {'theme':'base', 'themeVariables': { 'primaryColor':'#10b981','primaryTextColor':'#fff','primaryBorderColor':'#059669','lineColor':'#6366f1','secondaryColor':'#8b5cf6','tertiaryColor':'#ec4899'}}}%%
graph LR
    A["🌱 Foundation<br/><b>Lessons 1-3</b><br/><i>3-4 hrs</i>"] --> B["🔧 Intermediate<br/><b>Lessons 4-5</b><br/><i>5-6 hrs</i>"]
    B --> C["🚀 Advanced<br/><b>Lessons 6-7</b><br/><i>6-7 hrs</i>"]
    C --> D["💫 Multi-Agent<br/><b>Lessons 8-9</b><br/><i>10-12 hrs</i>"]
    D --> E["🏭 Production<br/><b>Lesson 10</b><br/><i>6-7 hrs</i>"]

    A -.-> A1["• Basic Conversational<br/>• Tool-Enhanced"]
    B -.-> B1["• State & Sessions<br/>• Async Streaming<br/>• Multi-modal"]
    C -.-> C1["• Hooks & Output<br/>• Context-Aware Tools<br/>• Conversation Mgmt"]
    D -.-> D1["• Graph, Swarm, Workflow<br/>• Agents-as-Tools<br/>• A2A Protocol"]
    E -.-> E1["• Guardrails & PII<br/>• Observability<br/>• Testing"]

    style A fill:#10b981,stroke:#059669,stroke-width:3px,color:#fff
    style B fill:#3b82f6,stroke:#2563eb,stroke-width:3px,color:#fff
    style C fill:#8b5cf6,stroke:#7c3aed,stroke-width:3px,color:#fff
    style D fill:#ec4899,stroke:#db2777,stroke-width:3px,color:#fff
    style E fill:#f59e0b,stroke:#d97706,stroke-width:3px,color:#fff
    style A1 fill:#d1fae5,stroke:#059669,stroke-width:2px,color:#065f46
    style B1 fill:#dbeafe,stroke:#2563eb,stroke-width:2px,color:#1e40af
    style C1 fill:#ede9fe,stroke:#7c3aed,stroke-width:2px,color:#6b21a8
    style D1 fill:#fce7f3,stroke:#db2777,stroke-width:2px,color:#9f1239
    style E1 fill:#fef3c7,stroke:#d97706,stroke-width:2px,color:#92400e

Loading

What you'll build: Simple chatbots → Tool-using agents → Stateful systems → Multi-agent teams → Production applications

Each lesson teaches distinct architectural patterns with working code you can run, modify, and build upon.

✨ Features

• 🚀 10 Comprehensive Lessons - Complete coverage of Strands SDK
• 🛠️ Hands-On Approach - Working code you can run immediately
• 📋 Clear Success Criteria - Know when you've mastered each concept
• 🔬 Built-in Experiments - Explore and modify examples
• 📚 Complete Documentation - Detailed explanations and troubleshooting
• 🏗️ Production-Ready Patterns - Safety, security, and observability
• ⚡ Modern Setup - Uses uv for fast Python package management

🏃‍♂️ Quick Start

Prerequisites

• Python 3.10+
• uv package manager (install guide)
• AI Model API Key - Choose one:
  • OpenAI (get key) - Recommended
  • Anthropic (get key) - Alternative
  • Google Gemini (get key) - Alternative
  • AWS Bedrock (setup guide) - Enterprise
  • Ollama (local, free) - For advanced users
• Basic understanding of Python async/await and decorators

Get Started in 4 Steps

Step 1: Clone the Repository

git clone https://github.com/jztan/strands-agents-learning.git
cd strands-agents-learning

Step 2: Install Dependencies

# Install all dependencies
uv sync

# ONE-TIME: Install Jupyter kernel (required for notebooks)
uv run python -m ipykernel install --user \
  --name=strands-learning \
  --display-name="Python (Strands)"

Step 3: Configure API Key

# Copy environment template
cp .env.example .env

# Edit .env and add ONE of the following:
# OPENAI_API_KEY=your_openai_key_here        # ← Recommended
# ANTHROPIC_API_KEY=your_anthropic_key_here  # ← Alternative
# GOOGLE_API_KEY=your_google_key_here        # ← Alternative

# Or configure AWS Bedrock (requires AWS credentials + model access):
# Step 1: Configure AWS credentials
#   aws configure  # Enter AWS Access Key ID and Secret Access Key
# Step 2: Enable model access (one-time)
#   Visit: https://console.aws.amazon.com/bedrock/
#   Go to: Model access (left sidebar)
#   Enable: Amazon Nova Lite model

Step 4: Run Your First Lesson

Option A: Jupyter Notebooks (interactive, recommended)

# Using Jupyter Notebook
uv run jupyter notebook lesson_01_hello_world.ipynb

# Or using Jupyter Lab (modern interface)
uv run jupyter lab lesson_01_hello_world.ipynb

Option B: Python Scripts (for reference)

uv run python lesson_01_hello_world.py

✨ That's it! You now have a working Strands agent with intelligent provider selection.

📓 Why Use Jupyter Notebooks?

Each lesson comes as both a Jupyter notebook (.ipynb) and Python script (.py). We recommend notebooks for learning:

• Run code cell-by-cell to understand each concept
• Experiment and modify inline without editing files
• See outputs immediately
• Add your own notes as markdown cells

Python scripts are provided for reference and production use.

📚 Learning Path & Agent Patterns

Level	Lesson	Agent Pattern	Topics
Foundation	Lesson 1: Hello World ✅	Basic Conversational	Agent basics, system prompts, sync/async
	Lesson 2: First Tool ✅	Tool-Enhanced	@tool decorator, calculator, error handling
	Lesson 3: Multiple Tools ✅	Tool-Enhanced	Tool coordination, weather/time/converter
Intermediate	Lesson 4: Agent State ✅	State & Sessions	agent.state API, FileSessionManager, persistence
	Lesson 5: Async, Executors & MCP ✅	Async Streaming, Multi-modal & Integration	Async tools, yield, executors, images, PDFs, MCP
Advanced	Lesson 6: Hooks & Output ✅	Hooks & Structured	Lifecycle hooks, Pydantic models, logging
	Lesson 7: Advanced Tools ✅	Context-Aware	Class-based tools, ToolContext, conversation mgmt
Multi-Agent	Lesson 8: Orchestration ✅	Graph, Swarm & Workflow	Multi-agent patterns, state sharing
	Lesson 9: Distributed Agents ✅	Agents-as-Tools & A2A	Hierarchical agents, cross-platform communication
Production	Lesson 10: Production ✅	Safety, Observability & Eval	Guardrails, PII redaction, OpenTelemetry, testing
Bonus	AWS Deployment 🚧	Lambda, EC2, ECS, Bedrock	Serverless, containers, managed AI, IaC

Progress: 10/10 lessons complete (100%) 🎉 • Each lesson includes: Learning objectives • Working code • Experiments • Success criteria • Common pitfalls

💡 Learning Tips

• Start Simple: Begin with Lesson 1, don't skip ahead
• Run the Code: Execute every example to see it in action
• Experiment: Modify examples and see what happens
• Use the Experiments Folder: Try your own variations
• Debug Thoughtfully: When things break, understand why

🛠️ Having issues? Check the Troubleshooting Guide

🤝 Contributing

We welcome contributions! Whether you want to report bugs, suggest improvements, or submit code - we'd love your help.

See CONTRIBUTING.md for:

• Development setup and workflow
• Coding standards and best practices
• Testing guidelines
• How to submit pull requests

📖 Additional Resources

• Strands Documentation - Official framework docs
• Strands GitHub - Source code and issues
• Strands Agents SDK: Three Things That Clicked - A practical walkthrough of building agents (by the author)

📄 License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

---

⬆️ Back to Top