OpenSearch Agentic Memory: Intelligent Context for AI Agents
summary
OpenSearch Agentic Memory describes a concrete system: a conversation memory index collects session logs, a background memory builder extracts facts via an LLM and resolves conflicts, then writes structured items into a long-term memory embeddings index (dense or sparse) with namespaced isolation. Exposes REST APIs and an MCB tool for LangChain/Strand integration, supports customizable prompts and BYO models, and the demos show cross-session personalization (shopping and travel agents remembering preferences and allergies) aiming for GA in 3.3.
AI-written summary · about
chapters (9)
- 0:00OpenSearch Agentic Memory intro
- 1:04AI agents memory challenges
- 2:54Solution with Agentic Memory
- 5:27Agentic Memory architecture
- 9:20Agentic Memory strategies
- 11:48Demo integration with frameworks
- 14:43Demo personalized recommendations
- 21:00Demo travel planning agent
- 26:31Q&A
related talks
21mPersonalize Search Results with OpenSearch Agentic Memory
26mOpenSearch: Conversations with Context
40mBuilding Intelligent Search Systems With Agentic RAG and OpenSearch
54mThe Anatomy of Memory in Humans and AI Agents
36mBuilding Next-gen Search With OpenSearch
30mBeyond Keywords: Building an AI-Powered Search That Reads Your Mind
37mAgentic AI Meets Search: Real-World Use Cases With...
51mSecrets of Agentic Memory Management