This repository provides a reference implementation for building generative AI financial services agents using Amazon Bedrock and LangChain. It targets developers looking to create personalized conversational agents for applications like chatbots and virtual assistants, enabling autonomous tool orchestration for human-like responses.

How It Works

The solution orchestrates a LangChain conversational agent powered by Anthropic's Claude 3 Sonnet on Amazon Bedrock. It leverages Amazon Lex for natural language understanding and an AWS Amplify website for the user interface. The agent utilizes Retrieval-Augmented Generation (RAG) with Amazon Kendra for authoritative data retrieval and Amazon DynamoDB for conversation memory, enabling context-aware, multi-turn interactions with source attribution.

Quick Start & Requirements

• Deployment requires AWS services including Amazon Bedrock, Amazon Lex, AWS Lambda, Amazon DynamoDB, Amazon Kendra, and AWS Amplify.
• Refer to the Deployment Guide for detailed setup instructions.

Highlighted Details

• Demonstrates Retrieval-Augmented Generation (RAG) for providing opinionated answers with source attribution.
• Utilizes DynamoDB for conversation memory, enabling context-aware responses.
• Supports chain-of-thought reasoning for agent decision-making.
• Integrates with Amazon Kendra for semantic search across various data sources.

Maintenance & Community

• This is an AWS sample solution, maintained by AWS.
• No specific community channels (Discord/Slack) are listed.

Licensing & Compatibility

• Licensed under the MIT-0 license.
• Permissive license suitable for commercial use and integration into closed-source applications.

Limitations & Caveats

The solution is presented as a launchpad and requires significant AWS service configuration and understanding for deployment and customization. Specific details on performance benchmarks or limitations of the underlying foundation models are not provided.