NirDiamant/Agent_Memory_Techniques logo

NirDiamant/Agent_Memory_Techniques

Free

Learn every agent memory technique for LLM agents.

FreeFree tier
Type
Open Source

About NirDiamant/Agent_Memory_Techniques

Agent memory is the set of techniques that let an LLM-based agent remember information across turns, sessions, and tasks. Without memory, an agent re-derives context every time and cannot personalize, learn, or maintain coherence over long interactions. This repository documents 30 distinct memory techniques, grouped into six families: short-term context management, long-term storage, cognitive architectures, retrieval and multi-agent patterns, batteries-included frameworks, and production deployment patterns. It includes runnable Jupyter notebooks covering conversation buffers, vector stores, knowledge graphs, episodic and semantic memory, working memory, MemGPT, Mem0, Letta, Zep, Graphiti, LoCoMo benchmarks, and production memory patterns.

Key Features

30 runnable Jupyter notebooks
Covers conversation buffer memory
Vector store memory
Knowledge graph memory
Episodic and semantic memory
Working memory
MemGPT integration
Mem0 integration
Letta integration
Zep integration

Pros & Cons

Pros
  • Comprehensive coverage of 30 distinct memory techniques
  • Runnable notebooks allow hands-on experimentation
  • Free and open source with no licensing costs
  • Integrates with popular frameworks like MemGPT, Letta, Zep, etc.
  • Structured learning paths for beginners and advanced users
Cons
  • Requires Jupyter environment to run notebooks
  • Educational resource, not a plug-and-play library
  • Assumes existing knowledge of LLM agent concepts
  • No direct API or hosted service; self-hosted execution needed

Best For

Learning and understanding agent memory techniques for LLMsBuilding LLM agents with persistent memoryImplementing production memory systems for AI agentsBenchmarking memory strategies with LoCoMo

FAQ

What is agent memory?
Agent memory is the set of techniques that let an LLM-based agent remember information across turns, sessions, and tasks. Without it, an agent cannot personalize, learn, or maintain coherence over long interactions.
How many memory techniques are covered?
The repository documents 30 distinct memory techniques, grouped into six families: short-term context management, long-term storage, cognitive architectures, retrieval and multi-agent patterns, batteries-included frameworks, and production deployment patterns.
Which third-party tools are integrated?
The notebooks cover integrations with MemGPT, Mem0, Letta, Zep, and Graphiti, among others.