This repository provides a comprehensive collection of proof-of-concept (PoC) code samples for leveraging Amazon Bedrock and generative AI across various use cases. It targets developers and researchers looking to quickly prototype and explore capabilities like natural language querying of data stores, multi-modal interactions, document processing, and conversational AI. The primary benefit is rapid experimentation with Amazon Bedrock's features through ready-to-run examples.

How It Works

The project is structured as a collection of independent Python and .NET projects, each demonstrating a specific generative AI application. Most Python samples utilize Streamlit for a user-friendly web interface, allowing direct interaction with Amazon Bedrock models. The .NET samples use Blazor for their frontends. The underlying approach involves integrating with Amazon Bedrock APIs to access various LLMs and multi-modal models, often combined with other AWS services like Amazon Transcribe, Kendra, or OpenSearch for data ingestion, retrieval, and specialized functionalities.

Quick Start & Requirements

• Python:
  • Install: pip install -r requirements.txt (within each sample directory).
  • Run: streamlit run app.py (within each sample directory).
  • Prerequisites: Amazon Bedrock access and CLI credentials, Python 3.10+. Specific samples may require additional AWS services (e.g., Kendra, RDS).
• .NET:
  • Install: Visual Studio with .NET 8.0.
  • Run: Launch the Blazor application from Visual Studio.
  • Prerequisites: Amazon Bedrock access and CLI credentials, .NET 8.0, Visual Studio. Access to Claude 3 Haiku model is noted.

Highlighted Details

• Extensive coverage of Amazon Bedrock features, including Converse API, Guardrails, Knowledge Bases (RAG), and model customization.
• Demonstrates multi-modal capabilities with Claude 3 for image analysis and generation.
• Includes samples for data querying across different relational databases (Athena, RDS, Redshift) and vector stores (OpenSearch Serverless).
• Showcases integration with other AWS services for tasks like speech-to-text, document summarization, and video chapter generation.

Maintenance & Community

The repository is part of the aws-samples organization, indicating official AWS backing. Contributions are welcome via standard GitHub pull requests.

Licensing & Compatibility

Licensed under the MIT-0 License, which permits unrestricted use, modification, and distribution, including for commercial purposes.

Limitations & Caveats

The repository contains "Proof of Concepts" (PoCs), implying that the code is illustrative and may require further development for production environments. Specific samples have dependencies on other AWS services that need to be provisioned and configured separately. Access to certain models, like Claude 3 Haiku, may require explicit approval.