Getting Started
Core Features
How-To Guides
Configuration
Troubleshooting
Integrations
This page explains how mem0 manages memory storage, focusing on recent simplifications to its configuration. mem0 uses specialized databases called vector stores to efficiently store and retrieve your memories. This ensures that your application can quickly find relevant information when needed.
The latest updates have streamlined the process of configuring these storage options, making it easier to set up and manage your memory system.
mem0, vector stores are used to store memories in a format that allows for fast and accurate semantic search.MemoryConfig: This is the main configuration object for your mem0 instance. It allows you to specify various settings, including how your memories are stored.VectorStoreConfig: A specific part of MemoryConfig that defines the settings for your chosen vector store.Previously, configuring your memory storage might have involved specifying a version parameter within your MemoryConfig. This parameter is no longer necessary and should be removed for a cleaner and more consistent setup.
MemoryConfigIf you were previously configuring your memory with a version parameter alongside VectorStoreConfig, you should now remove it. The system automatically handles the correct versioning.
# Before (if you had a version parameter with VectorStoreConfig)
config = MemoryConfig(
version="v1.1", # This parameter is no longer needed here
vector_store=VectorStoreConfig(...)
)
mem0 has improved the reliability and functionality of its vector stores, including fixes for OpenSearch and general enhancements across all supported storage options.
Here's a simple example demonstrating the updated way to configure your memory with a vector store, without the deprecated version parameter:
from mem0 import Memory, MemoryConfig
# Assuming VectorStoreConfig is imported or defined elsewhere
# from mem0.configs import VectorStoreConfig
# Configure your memory with a vector store
# (Details of VectorStoreConfig(...) depend on your specific vector store)
my_config = MemoryConfig(
vector_store=VectorStoreConfig(
# ... your specific vector store settings here ...
)
)
# Initialize Memory with your configuration
memory = Memory(
If you encounter deprecation warnings related to version="v1.0" in your configuration, ensure you remove any explicit version parameters from your MemoryConfig initialization.