Note: This repository is archived and no longer maintained. Chatting with a proprietary knowledge base via retrieval-augmented generation is now general practice across every agentic solution (ChatGPT, Copilot, and the like), so there is no need to keep this project evolving. It remains here as a record of the original work.
ChatLibrary is a generative AI-powered chat application designed to answer user questions based on a proprietary knowledge base. It utilizes advanced language models and vector databases to provide accurate and context-aware responses along with references and user feedback collection.
- Clone the repository:
git clone https://github.com/rnepal2/ChatLibrary.git cd ChatLibrary - Install the required packages:
pip install -r requirements.txt
- Set up environment variables:
export AZURE_OPENAI_API_KEY='your_api_key' export AZURE_OPENAI_ENDPOINT='your_endpoint'
Run the application using: streamlit run app.py --server.port 3000
- Streamlit: For building the web application interface.
- LangChain: For managing language model interactions.
- Azure OpenAI: For leveraging advanced language models.
- ChromaDB: For vector database management.
