Generative AI  

Empowering Agentic AI with Persistent Memory using Strands Agents SDK

Introduction

Strands Agents is an open-source SDK that simplifies the development of AI agents capable of using tools, making decisions, and automating workflows — moving far beyond basic chatbot interactions.

In this article, you will learn how to create a travel buddy agent that remembers user preferences using the Strands Agents SDK and FAISS as the local vector store backend.

Strands Agents SDK includes a built-in memory tool called Mem0, which supports multiple backends like FAISS (local), OpenSearch (AWS-native), and the Mem0 Platform (API-based). This memory layer enables agents to store, retrieve, and reason with contextual knowledge — forming the foundation for stateful and adaptive user experiences.

In this example, you will use FAISS (via the faiss-cpu package), which is the default backend for local development. No additional configuration is required to get started.

Pre-requisites

  1. Install or update to the latest version of the AWS CLI.
  2. Get credentials to grant programmatic access.
  3. Visual Studio Code.
  4. Python 3.10+
  5. Access to Amazon Bedrock foundation model. The default model provider is Amazon Bedrock and the default model is Claude 3.7 Sonnet in the US Oregon (us-west-2) region.

Build the Travel Buddy Agent using Strands Agents SDK

Perform the following steps to create and configure an agent using Strands Agents SDK.

  1. Open Visual Studio Code.
  2. Navigate to the folder where you want to create your Python file.
  3. Open a new PowerShell terminal in Visual Studio Code.
  4. Run the following command to create a virtual environment. 
    python -m venv .venv
  5. Run the following command to activate the virtual environment. 
    .venv\Scripts\Activate.ps1
  6. Run the following command to install the required packages. 
    pip install strands-agents faiss-cpu
  7. Create a new Python file and name it as travel_agent.py.
  8. Copy and paste the below code to travel_agent.py. 
    from strands import Agent
from strands_tools import mem0_memory

# Sample user ID for demo purpose
user_id = "user_123"

# Define the system prompt for Travel Buddy agent
system_prompt = """
You are Travel Buddy, a friendly AI assistant helping users plan vacations.
You remember their preferences — such as favorite destinations, travel class, budget, and dietary restrictions.
Use that memory to suggest trips that match their style.
"""

# Initialize the Travel Buddy agent with memory capability (FAISS- Default)
agent = Agent(
    tools=[mem0_memory],
    system_prompt=system_prompt,
)

# Store user's travel preferences as memory
agent.tool.mem0_memory(
    action="store",
    content="User prefers tropical destinations like Bali and Maldives, Business class flights, vegetarian food, and a budget under $2500.",
    user_id=user_id
)

# Ask a question that utilizes stored memory
print("\n Travel Buddy Recommendation:\n")
response = agent("Where should I travel next?", user_id=user_id)

# List all stored memories for the user
print("\n Stored Memories:\n")
memories = agent.tool.mem0_memory(
    action="list",
    user_id=user_id
)

# Retrieve specific memory about destination preference
destination_memory = agent.tool.mem0_memory(
    action="retrieve",
    query="What kind of destinations do I prefer?",
    user_id=user_id
)
print("\n Retrieved Destination Preference:\n", destination_memory)
  9. Run the following command to execute your Python code. 
    python -u .\travel_agent.py

Output

Build the Travel Buddy Agent using Strands Agents SDK

Summary

In this article, you learned how to build an Agentic AI assistant — a Travel Buddy — using the Strands Agents SDK and FAISS-based memory.

You explored how to:

  • Configure an agent with memory
  • Store and retrieve user preferences
  • Deliver personalized recommendations based on memory

This foundational memory capability sets the stage for building adaptive, stateful, and human-like AI assistants.

Membership not found