MemGPT
OS-inspired memory management for LLMs enabling infinite context and persistent agents
MemGPT
OS-inspired memory management for LLMs enabling infinite context and persistent agents
MemGPT is a research project and open-source system that implements operating system-inspired memory management for large language models, enabling them to maintain persistent state across conversations and manage effectively unlimited context through tiered storage. By giving LLMs the ability to read from and write to external memory stores, MemGPT enables agents that accumulate knowledge over time and recall relevant information from past interactions. The academic paper and accompanying code have influenced subsequent approaches to agent memory, including the Letta platform that continued the MemGPT research.
Key Features
- ✓OS-inspired memory
- ✓Infinite context
- ✓Persistent state
- ✓External storage
- ✓Research paper
Quick Info
- Category
- AI Agents
- Pricing
- Free
More AI Agents Tools
Vertex AI Agent Builder
AI Agents & AutomationGoogle Cloud's platform for enterprise AI agent development
Amazon Bedrock Agents
AI Agents & AutomationBuild autonomous AI agents using AWS foundation models
Riku AI
AI Agents & AutomationNo-code platform for building and deploying AI models
Pabbly Connect
AI Agents & AutomationAffordable AI automation platform for small businesses