rag-with-amazon-bedrock-and-pgvector and rag-with-amazon-bedrock-and-opensearch
These two tools are competitors, as both are opinionated sample implementations for building a RAG application with Amazon Bedrock, but they differ in their choice of vector database backend, with one using PGVector and the other OpenSearch.
About rag-with-amazon-bedrock-and-pgvector
aws-samples/rag-with-amazon-bedrock-and-pgvector
Opinionated sample on how to build/deploy a RAG web app on AWS powered by Amazon Bedrock and PGVector (on Amazon RDS)
Implements an event-driven PDF ingestion pipeline using S3 Event Notifications and Lambda functions to extract text, generate OpenAI embeddings, and populate PGVector for semantic search. The full stack uses AWS CDK for infrastructure-as-code, deploys a Streamlit frontend via ECS with Application Load Balancer routing, and secures access through Cognito authentication while integrating LangChain for RAG orchestration across Bedrock foundation models.
About rag-with-amazon-bedrock-and-opensearch
aws-samples/rag-with-amazon-bedrock-and-opensearch
Opinionated sample on how to build and deploy a RAG application with Amazon Bedrock and OpenSearch
Implements an event-driven document ingestion pipeline where PDFs trigger Lambda functions to extract text, generate vector embeddings via OpenAI, and automatically index content in OpenSearch for semantic search. The architecture layers Streamlit frontend, LangChain orchestration, and Bedrock foundation models on ECS with Cognito authentication and ALB routing, using AWS CDK for infrastructure-as-code deployment across multiple stacks.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work