AWS and Mem0 Partner to Bring Persistent Memory to Next-Gen AI Agents with Strands

We’re excited to announce a strategic partnership between AWS and Mem0 to enhance the capabilities of agentic AI. This collaboration integrates Mem0’s advanced memory layer into Strands Agents SDK, bringing persistent, contextual, and highly personalized memory to cloud-native autonomous agents.

Introducing Strands + Mem0: A New Standard for Agentic Intelligence

The Strands Agents SDK is a lightweight, production-ready framework for building AI agents with built-in tools, multi-agent support, and seamless integration with Amazon Bedrock. It offers full observability, model-agnostic flexibility, and scalable deployment across AWS Lambda, Fargate, and EC2. By integrating Mem0, Strands supports a persistent memory layer that allows agents to retain and learn from past interactions - transforming stateless LLM-based tools into intelligent, evolving agents.

Why Memory Matters for AI Agents

Large language models (LLMs) have revolutionized how we build intelligent systems, but they remain inherently stateless. This limitation restricts their ability to retain information across interactions, learn user preferences, or build context over time. This often leads to repetitive, disconnected user experiences.

Mem0 addresses this limitation by offering a production-grade, memory layer for AI agents. With Mem0, developers can build agents that:

• Personalize behavior and adapt to individual users
• Maintain long-term reasoning across conversations
• Minimize redundant queries and reduce token costs
• Retain and evolve knowledge across sessions

How Mem0 Enhances Strands

Mem0 is an open-source, production-grade memory layer that combines multiple layers of intelligence:

• Semantic memory retrieval via vector embeddings
• Graph-based memory (Mem0ᵍ) for understanding complex relationships
• LLM-assisted memory updates that automatically deduplicate, merge, or revise memory content

This enables Strands agents to go beyond static responses and instead reason, reflect, and evolve based on historical interactions and structured knowledge.

Practical Guide: Adding Memory to a Strands Agent

Mem0 integrates seamlessly with Strands in just a few lines of code. Each interaction is scoped by user_id, ensuring context stays both personalized and securely isolated. Here's a simple example:

import os
from strands import Agent
from strands_tools import mem0_memory, use_llm

# Define user identity
user_id = "Alex"

# Create a memory-enabled Strands agent
agent = Agent(
    system_prompt="You are a helpful assistant that remembers user preferences.",
    tools=[mem0_memory, use_llm],
)

# Store a fact
agent("I live in San Francisco.", user_id=user_id)

# Ask a follow-up that uses memory
agent("What's the weather like today?", user_id=user_id)

# Ask another contextual question
agent("What about tomorrow?", user_id=user_id)

# Check what the agent remembers
agent("What do you remember about me?", user_id=user_id)

Check out this example on GitHub.

Real-World Use Cases

The combination of Strands and Mem0 unlocks powerful applications across industries:

• Personalized Healthcare Assistants: Healthcare Agents that remember patient histories, conditions, and treatment preferences, supporting more effective and continuous care.
• E-Commerce Recommendations: E-commerce Agents that recall past purchases, preferences, and browsing behavior to deliver highly personalized product suggestions, increasing sales conversion rates.
• Customer Support Automation: Customer Support Agents that retain customer history and issue resolution records, improving customer satisfaction and reducing support overhead.
• Enterprise Knowledge Management: Internal copilots that learn company processes, tools, and workflows, retaining institutional knowledge and assisting across departments.

Why This Partnership Matters

The AWS-Mem0 partnership is a significant step forward in the evolution of agentic AI. This integration enables developers to:

• Build agents that deliver highly personalized experiences, fostering stronger user engagement.
• Optimize performance and reduce costs with Mem0’s efficient memory management.
• Scale securely on AWS infrastructure with native support for Bedrock models

Together, we’re setting a new bar for what intelligent agents can do in enterprise and developer ecosystems.

Get Started Today

Start building memory-enhanced AI agents with Strands and Mem0:

1. Explore the Strands Agents SDK
2. Strands Agents with Mem0
3. Sign up for Mem0
4. Explore the Mem0 docs

As intelligent agents become core to how businesses interact, automate, and operate, memory will be a defining capability. AWS and Mem0 are committed to empowering developers with the tools to build agents that don’t just respond - they remember, learn, and evolve.