This project provides a chat interface for interacting with uploaded documents (PDF, DOCX, TXT) using OpenAI's ChatGPT and the LangChain framework. It's designed for users who need to query and extract information from their own documents via a conversational AI.

How It Works

The application leverages the T3 Stack (Next.js, Tailwind CSS, tRPC) for its frontend and backend. Documents are processed and vectorized using LangChain, then stored and indexed in Weaviate, an open-source vector database. Cloudflare R2 is used for object storage. Users interact through a Mantine UI-based chat interface, with Zustand managing application state.

Quick Start & Requirements

• Install dependencies: npm install
• Configure environment variables in .env (Weaviate host/key, Cloudflare R2 credentials, OpenAI API key).
• Start the application: npm run dev
• Prerequisites: Node.js v18+, OpenAI API key, Weaviate instance, Cloudflare R2 account.

Highlighted Details

• Utilizes Weaviate for efficient vector storage and retrieval.
• Employs Cloudflare R2 for scalable object storage.
• Built with the T3 Stack for a robust full-stack experience.
• Features a user-friendly chat interface powered by Mantine UI.

Maintenance & Community

This project was a submission for a Bellingcat hackathon. Team members include Radu Ciocan (code) and Ana State (design). Further community or maintenance information is not detailed in the README.

Licensing & Compatibility

The repository does not explicitly state a license in the provided README. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

The README does not specify supported document types beyond PDF, DOCX, and TXT. There is no mention of performance benchmarks, scalability considerations, or error handling strategies. The project's long-term maintenance status is unclear.