OpenSearch Flow and Agent Frameworks: Easy AI
summary
OpenSearch 2.13 adds a flow framework and 3.0 ships workflow templates plus an agent framework that runs tool chains; the demo uses the neural plugin (ingest and search pipelines), connector-model IDs (Bedrock, SageMaker), and a ~4M‑document Amazon PQA dataset. The opensearch-ml-quickstart automates connector/model creation, index/pipeline deployment and examples for dense HNSW, sparse, hybrid, conversational and agentic flows; lexical and semantic demos return sensible results, and the conversational agent takes ~30s per query, traces its tool calls and produces usable summaries despite a minor output-printing bug.
AI-written summary · about
chapters (9)
- 0:00OpenSearch and AI search intro
- 2:53Hybrid conversational search
- 8:28AI search with neural plugin
- 17:24AI search deployment challenges
- 18:28OpenSearch ML Quick Start
- 22:05Lexical search demo
- 25:18Semantic search demo
- 30:41Agentic conversational demo
- 41:41Summary and ML benefits
related talks
35mFrom Semantic Search to AI Agents: Quick Start on Production-Ready ML Appl...
27mElevating AI Applications with OpenSearch's Flow Framework and RAG Tool
25mKeynote: OpenSearch in the Agentic Era
43mOpenSearch & AI Driven Cloud Native Platforms: Adaptive & Agentic Microserv...
26mOpenSearch: Conversations with Context
06mSponsored Keynote: OpenSearch - See Everything: Open Observability for Agentic AI
42mOpenSearch With Gaia: Building Next-Gen AI-Driven Search and Analytics Workflows
30mOpenSearch for GenAI Storage: AgentScope Studio Practice