​
Available Components

AgentMemory_Memory Node

Specialized memory for agent-based conversations

Buffer Window Memory Node

Maintains a fixed number of recent message exchanges

Buffer Memory Node

Simple, short-term memory for recent interactions

Conversation Summary Memory Node

Handles long-running conversations where full history is too large

Conversation Summary Buffer Memory Node

Combines recent messages with a summary of older context

DynamoDB Chat Memory Node

AWS-based persistent storage for chat history

MongoDB Atlas Chat Memory Node

Persistent storage of chat history in MongoDB Atlas

Redis-Backed Chat Memory

Similar to Upstash, but with standard Redis

Upstash Redis-Backed Chat Memory

Fast, scalable memory storage using Upstash Redis

ZepMemory Node

Long-term memory storage using Zep open-source memory server

ZepMemoryCloud

Cloud-based version of ZepMemory for scalable, hosted solutions

​
Selection Guide

Choose the appropriate memory component based on your needs:
  1. Simple Interactions
    • Use Buffer Memory or Buffer Window Memory for short conversations
    • Best for basic chatbots and brief interactions
  2. Long Conversations
    • Use Conversation Summary Memory or Conversation Summary Buffer Memory
    • Ideal when you need to maintain context over extended dialogues
  3. Persistent Storage
    • MongoDB Atlas: For scalable, cloud-based document storage
    • Redis/Upstash Redis: For fast, in-memory data access
    • DynamoDB: For AWS-integrated applications
    • Choose based on your existing infrastructure and scaling requirements
  4. Specialized Use Cases
    • AgentMemory: For complex agent-based systems
    • ZepMemory/ZepMemoryCloud: For managed, scalable memory solutions
    • Upstash Redis: For serverless and edge deployments