Agent_Memory_Techniques

Why we included this project

Most agent-memory tutorials stop at a demo and never show you what to do next. This repository actually does the work for you: 30 runnable Jupyter notebooks that start with the basic conversation buffer and climb through vector stores, knowledge graphs, and episodic or semantic memory before reaching production-oriented frameworks like MemGPT, Mem0, Letta, Zep, and Graphiti. Because every technique ships as an executable notebook, it works as a structured course for newcomers and a quick reference for anyone who just needs to solve today's memory problem. The bundled learning paths and decision tree help you pick the right approach for a single chat session, cross-session persistence, or a full multi-store architecture, and the final notebooks on evaluation, benchmarking, and deployment cover the part most guides skip: how to measure and ship a memory layer, not just build one. That makes it a genuinely useful way to compare memory designs before you commit to one in a real product.