Developer Builds Terraform Module to Automate Amazon Bedrock RAG Pipeline End-to-End

A developer has published a Terraform module that automates the full infrastructure setup for a Retrieval-Augmented Generation (RAG) 'Chat with PDF' application on Amazon Bedrock. The module provisions an S3 bucket, OpenSearch Serverless vector database, a Bedrock Knowledge Base, and two Lambda functions — one for ingestion and one for querying — with a single terraform apply command. The project was motivated by the absence of any existing GitHub module covering the complete end-to-end pipeline after Amazon Bedrock launched in April 2023. Documents uploaded to S3 are automatically chunked, embedded via Amazon Titan, and indexed into OpenSearch, while a query Lambda uses Bedrock's RetrieveAndGenerate API to return answers with source citations. The module is structured across four sub-modules for storage, OpenSearch, Bedrock, and Lambda, with a planned third part to add API Gateway in front of the query function.