🧠 Tutorial 5.1: In-Memory Conversation Agent

Welcome to your first step into session management! This tutorial teaches you how to create an AI agent that can remember conversations within a single session using InMemorySessionService.

🎯 What You'll Learn

• InMemorySessionService: Basic session management for temporary conversations
• Session Creation: How to create and manage conversation sessions
• State Management: Storing and retrieving conversation context
• Event Tracking: Recording conversation history
• Multi-turn Conversations: Building agents that remember context

🧠 Core Concept: In-Memory Sessions

InMemorySessionService stores session data in your computer's RAM (memory). This means:

• ✅ Fast access - No database queries needed
• ✅ Simple setup - No external dependencies
• ❌ Temporary storage - Data is lost when the program stops
• ❌ No persistence - Can't remember across program restarts

Perfect for:

• Development and testing
• Temporary conversations
• Prototyping memory features
• Single-session applications

🔧 Key Components

1. InMemorySessionService

from google.adk.sessions import InMemorySessionService

2. Session Lifecycle

┌─────────────┐    ┌─────────────┐    ┌─────────────┐
│   CREATE    │───▶│   USE       │───▶│   CLOSE     │
│  SESSION    │    │  SESSION    │    │  SESSION    │
└─────────────┘    └─────────────┘    └─────────────┘

3. Session Data Structure

{
    "session_id": "unique_session_id",
    "user_id": "user_identifier", 
    "state": {
        "conversation_history": [...],
        "user_preferences": {...},
        "current_context": "..."
    },
    "events": [
        {"type": "user_input", "content": "...", "timestamp": "..."},
        {"type": "agent_response", "content": "...", "timestamp": "..."}
    ]
}

🚀 Tutorial Overview

In this tutorial, we'll create a Personal Assistant Agent that:

• Remembers your name and preferences
• Tracks conversation history
• Provides personalized responses
• Demonstrates basic session management

📁 Project Structure

5_1_in_memory_conversation/
├── README.md              # This file - concept explanation
├── requirements.txt       # Dependencies
├── agent.py              # Main agent with session management
└── app.py                # Streamlit web interface

🎯 Learning Objectives

By the end of this tutorial, you'll understand:

• ✅ How to create and manage sessions
• ✅ How to store and retrieve conversation state
• ✅ How to track conversation events
• ✅ How to build multi-turn conversations
• ✅ Basic session lifecycle management

🚀 Getting Started

1. Install dependencies:

   pip install -r requirements.txt
   

2. Set up your environment:

   # Create a .env file with your Google AI API key
   echo "GOOGLE_API_KEY=your_api_key_here" > .env
   

3. Run the agent:

   # Start the Streamlit app
   streamlit run app.py
   

4. Test the memory:

   • Tell the agent your name: "My name is John"
   • Ask about your preferences: "What do you know about me?"
   • Have a conversation and see how it remembers context

🔍 Code Walkthrough

Key Session Management Code:

# 1. Create session service
session_service = InMemorySessionService()

# 2. Create a new session
session = await session_service.create_session(
    app_name="personal_assistant",
    user_id="user123"
)

# 3. Update session state
await session_service.update_session_state(
    session_id=session.session_id,
    state={"user_name": "John", "preferences": ["travel", "music"]}
)

# 4. Add events to track conversation
await session_service.add_event(
    session_id=session.session_id,
    event_type="user_input",
    content="My name is John"
)

🎯 Testing Your Agent

Try these conversation flows to test memory:

Flow 1: Personal Information

User: "My name is Alice"
Agent: "Nice to meet you, Alice! How can I help you today?"

User: "What's my name?"
Agent: "Your name is Alice! I remember you told me that."

Flow 2: Preferences

User: "I love pizza and hiking"
Agent: "Great! I'll remember that you love pizza and hiking."

User: "What are my interests?"
Agent: "Based on our conversation, you love pizza and hiking!"

Flow 3: Context Continuity

User: "I'm planning a trip"
Agent: "That sounds exciting! Since you mentioned hiking, would you like recommendations for hiking destinations?"

User: "Yes, where should I go?"
Agent: "Given your love for hiking, I'd recommend..."

🔗 Next Steps

After completing this tutorial, you'll be ready for:

• Tutorial 5.2: Persistent Conversation - Learn database-based session storage
• Tutorial 5.3: Cloud Memory - Explore cloud-based memory solutions

💡 Pro Tips

• Test Multi-turn Conversations: Have extended conversations to see memory in action
• Monitor Session State: Use the web interface to inspect what the agent remembers
• Experiment with State: Try storing different types of data in the session state
• Understand Limitations: Remember that in-memory sessions are temporary

🚨 Important Notes

• Data Loss: In-memory sessions are lost when you restart the application
• Single Process: Sessions only work within the same Python process
• Memory Usage: Large conversation histories will consume RAM
• Development Only: Use in-memory sessions for development, not production