aws-solutions/generative-ai-application-builder-on-aws

Generative AI Application Builder on AWS facilitates the development, rapid experimentation, and deployment of generative artificial intelligence (AI) applications without requiring deep experience in AI. The solution includes integrations with Amazon Bedrock and its included LLMs, such as Amazon Titan, and pre-built connectors for 3rd-party LLMs.

What it solves

This project provides a web-based management dashboard that allows users to rapidly deploy, experiment with, and productionize customizable Generative AI use cases on AWS. It eliminates the need for manual infrastructure setup for every new AI application, enabling users to compare different Large Language Models (LLMs) and integrate them into production applications once optimized.

How it works

The solution consists of a central Deployment Dashboard and individual Use Case stacks:

  • Deployment Dashboard: A web UI (hosted on S3/CloudFront) that allows administrators to configure and launch AI use cases. It uses AWS Lambda and CloudFormation to automate the creation of the necessary AWS resources.
  • Text Use Cases: Implements a chatbot-style interface using LangChain to orchestrate requests. It can integrate with Amazon Kendra or Knowledge Bases for Amazon Bedrock for RAG (Retrieval-Augmented Generation) capabilities, querying enterprise data before sending the prompt to an LLM hosted on Amazon Bedrock or Amazon SageMaker AI.
  • Agent Use Cases: Deploys an interface that connects to Amazon Bedrock Agents to fulfill specific user intents.
  • Infrastructure: Uses Amazon API Gateway (WebSockets for streaming), Amazon Cognito for authentication, and Amazon DynamoDB for session and configuration management.

Who it’s for

It is designed for novice to experienced users—including DevOps engineers, administrators, and business users—who want to quickly prototype and scale Generative AI applications within the AWS ecosystem.

Highlights

  • Rapid Experimentation: Ability to deploy and compare multiple LLM combinations with metric tracking via Amazon CloudWatch.
  • Enterprise Ready: Supports company-specific data integration for real-world business problems.
  • Flexible LLM Providers: Integrates with both Amazon Bedrock and Amazon SageMaker AI.
  • Modular Architecture: Built with nested CloudFormation stacks for easy extensibility.
  • RAG Support: Built-in integration for querying enterprise knowledge bases.