Where does mem0 save memory if not set vector database?

[hongbo-miao]

I saw mem0 supports vector database for memory storage: https://docs.mem0.ai/components/vectordbs/overview.

For example,

config = {
    "vector_store": {
        "provider": "qdrant",
        "config": {
            "collection_name": "test",
            "host": "localhost",
            "port": 6333,
        }
    }
}

I am wondering when not vector_store field, where does mem0 save memory if not set vector database? Is it just temporally saving in the server memory? Thanks!

config = {
    "llm": {
        "provider": "openai",
        "config": {
            "model": config.MEMORY_MODEL,
            "temperature": config.MEMORY_MODEL_TEMPERATURE,
            "max_tokens": config.MEMORY_MODEL_MAX_TOKENS,
            "api_key": config.OPENAI_API_KEY,
        },
    },
    "embedder": {
        "provider": "openai",
        "config": {
            "api_key": config.OPENAI_API_KEY,
        },
    },
}

[parshvadaftari]

Hey @hongbo-miao that's a great question, since qdrant is used default, mem0 stores locally at /tmp/qdrant by default if novector_store config is provided.

[hongbo-miao]

I see, thanks for the info!

[azarboon]

I believe this response is incorrect or misleading. For example, in the following code, which is taken from the Strands Agents example, no config is defined. Apparently, this results in the data being stored only in memory, because each time the application is restarted, all previously stored data is lost. This indicates that persistence is not enabled by default. Please clarify this behavior and explain under what conditions, if any, data is persisted beyond the process lifecycle.

import os
import logging
from dotenv import load_dotenv

from strands import Agent
from strands_tools import mem0_memory, use_llm

logger = logging.getLogger(__name__)

# Load environment variables from .env file if it exists
load_dotenv()

USER_ID = "mem0_user"

# System prompt for the memory agent
MEMORY_SYSTEM_PROMPT = f"""You are a personal assistant that maintains context by remembering user details.

Capabilities:
- Store new information using mem0_memory tool (action="store")
- Retrieve relevant memories (action="retrieve")
- List all memories (action="list")
- Provide personalized responses

Key Rules:
- Always include user_id={USER_ID} in tool calls
- Be conversational and natural in responses
- Format output clearly
- Acknowledge stored information
- Only share relevant information
- Politely indicate when information is unavailable
"""

# Create an agent with memory capabilities
memory_agent = Agent(
    system_prompt=MEMORY_SYSTEM_PROMPT,
    tools=[mem0_memory, use_llm],
)

# Initialize some demo memories
def initialize_demo_memories():
    """Initialize some demo memories to showcase functionality."""
    content = """My name is Alex. I like to travel and stay in Airbnbs rather than hotels. I am planning a trip to Japan next spring. I enjoy hiking and outdoor photography as hobbies. I have a dog named Max. My favorite cuisine is Italian food."""  # noqa
    memory_agent.tool.mem0_memory(action="store", content=content, user_id=USER_ID)

# Example usage
if __name__ == "__main__":
    print("\n🧠 Memory Agent 🧠\n")
    print("This example demonstrates using Strands Agents' memory capabilities")
    print("to store and retrieve information.")
    print("\nOptions:")
    print("  'demo' - Initialize demo memories")
    print("  'exit' - Exit the program")
    print("\nOr try these examples:")
    print("  - Remember that I prefer window seats on flights")
    print("  - What do you know about me?")
    print("  - What are my travel preferences?")
    print("  - Do I have any pets?")

    # Interactive loop
    while True:
        try:
            user_input = input("\n> ")

            if user_input.lower() == "exit":
                print("\nGoodbye! 👋")
                break
            elif user_input.lower() == "demo":
                initialize_demo_memories()
                print("\nDemo memories initialized!")
                continue

            # Call the memory agent
            memory_agent(user_input)

        except KeyboardInterrupt:
            print("\n\nExecution interrupted. Exiting...")
            break
        except Exception as e:
            print(f"\nAn error occurred: {str(e)}")
            print("Please try a different request.")