AI Chat Web App with Bedrock RAG

A simple web application that provides Generative AI chat functionality using Amazon Bedrock with Retrieval-Augmented Generation (RAG) capabilities through Bedrock Knowledge Base.

Overview

This project implements a conversational AI interface that can retrieve information from a knowledge base to provide more accurate and contextual responses. The application consists of:

• Frontend: Web interface hosted on AWS Amplify using ReactJS
• Backend: API Gateway and Lambda function (Python)
• AI Services: Amazon Bedrock for FM access and Knowledge Base for RAG capabilities
• Vector Database: Amazon OpenSearch Serverless
• Data Source: Amazon S3 for data source storage

Architecture Diagram

Project Structure

claude-chat-app/
├── amplify/                                    # AWS Amplify configuration and backend code
│   └── backend/
│       └── function/                           # Lambda function code
│           └──bedrockRagFunction/src/index.py  # Edit the knowledge base ID in this file
├── prerequisites/
│   ├── data/                                   # Local data source file for knowledge base   
│   ├── create_bedrock_kb.py                    # Script to create Bedrock Knowledge Base
│   ├── requirements.txt                        # Python modules needed
│   └── README.md                               # Instructions for the prerequisites
├── src/
│   ├── App.js                                  # Main application component
│   ├── App.css                                 # Application styles
│   └── index.js                                # Application entry point
├── package.json                                # NPM dependencies
└── README.md                                   # Project documentation (this README file)

Getting Started

Clone the Repository

# Clone this repository
git clone https://github.com/MardiantoS/chat-ai-bedrock-rag.git

# Navigate to the project directory
cd chat-ai-bedrock-rag

Prerequisites

• AWS Account with appropriate permissions
• AWS CLI configured with access and secret keys
• AWS Amplify package
• Python 3.9 or higher
• Node.js and npm (for the frontend)

Setup Instructions

1. Knowledge Base Setup

Before deploying the main application, you need to set up the Knowledge Base:

1. Navigate to the prerequisites directory
2. Follow the README instructions within the directory
3. Note the Knowledge Base ID that is generated from the script
4. Edit amplify/backend/function/bedrockRagFunction/src/index.py and update the Knowledge Based ID placeholder with the one generated in the previous step

2. Install all Node packages

1. Install Amplify CLI if you haven't already

npm install -g @aws-amplify/cli

2. Install dependency packages:

npm install

3. Setup the backend: API Gateway and Lambda

1. Deploy the backend services:

amplify push

4. Run test locally (optional, but recommended)

1. Run the app in the development mode.

npm start

2. Open http://localhost:3000 to view it in your browser.

5. Frontend Deployment with AWS Amplify

Once your local test is successful, you can deploy the frontend. If you're running it for the first, run

• Add hosting to your Amplify project:

  amplify add hosting

• Choose your hosting options:

  ? Select the plugin module to execute: Hosting with Amplify Console
  ? Choose a type: Manual deployment
  

• Deploy your application:

  amplify publish

Usage

1. Navigate to the deployed Amplify URL in your web browser
2. Start a conversation in the chat interface
3. You can ask questions about JPMorgan Cahse shareholder and financial highlight reports for 2023 and 2024, and the application will use Retrieval-Augmented Generation (RAG) capability from the Amazon Bedrock Knowledge Base to provide more accurate and contextual responses

Sample Questions

Here are some example questions you can ask:

• "What was net income per share in 2022, 2023, and 2024?"
• "What are the significant challenges in 2024?"
• "When did Jamie Dimon start as CEO of BankOne?"

License

This project is licensed under the MIT License - see the LICENSE file for details.

Citations

If you use this code in your projects or research, please include the following citation:

Mardianto Hadiputro. (2025). AI Chat Application with Amazon Bedrock RAG. GitHub. https://github.com/MardiantoS/chat-ai-bedrock-rag

Acknowledgments

• The prerequisite script is based on the RAG examples from Amazon Bedrock Workshop | Knowledge Bases and RAG