Elevating AI Applications with OpenSearch's Flow Framework and RAG Tool
summary
OpenSearch 2.13 adds a Flow Framework that automates the full RAG pipeline: it creates connectors to embedding and LLM endpoints, registers models, builds ingest pipelines and indexes, and deploys a RAG agent with one API call. The demo wires an AWS Bedrock Titan embedding model plus a Cloudy LLM, ingests conference text with embeddings, runs a vector neural query, and returns an answer citing document provenance. A single deprovision API removes all created resources. This is a concrete one-click RAG builder, not just a conceptual overview, and it works end to end as shown. RAG Tool no-code/low-code
AI-written summary · about
chapters (14)
- 0:00Introduction to speakers
- 0:42What is OpenSearch
- 1:46Retrieval Augmented Generation basics
- 3:30RAG tool in OpenSearch
- 6:01Current ML setup for conversation search
- 7:27Introducing Flow Framework
- 9:35Live demo setup
- 13:09Demo creating connectors
- 14:55Demo using RAG tool
- 16:21Flow Framework agent demo
- 17:43Flow Framework use case demo
- 19:55Deprovisioning resources
- 20:35Sample use cases and roadmap
- 21:37Q&A session
related talks
16mOpenRAG: An open-source stack for RAG
40mBuilding Intelligent Search Systems With Agentic RAG and OpenSearch
43mOpenSearch Flow and Agent Frameworks: Easy AI
38mA Visual AI Search Flow Builder for OpenSearch: Build GenAI and Vector...
25mSet Up an Open Source RAG in 30 Minutes
38mAI-Ready Data Foundation Using Amazon OpenSearch
22mOrchestrating Multimodal RAG With Agentic Workflows With OpenSearch
36mBuilding Next-gen Search With OpenSearch