This project provides a document chatbot that allows users to upload files and converse with them using GPT and vector embeddings. It's built for users who want to interact with their documents in a conversational manner, leveraging technologies like LangchainJS and Supabase for efficient data retrieval and processing.

How It Works

The chatbot utilizes LangchainJS to process uploaded documents, breaking them down into manageable chunks. These chunks are then converted into vector embeddings and stored in a Supabase vector database. When a user asks a question, the system converts the query into an embedding, searches the Supabase database for relevant document chunks, and then feeds these chunks along with the query to a GPT model for a coherent answer.

Quick Start & Requirements

• Install: npm install
• Run: npm run dev
• Prerequisites: Node.js, Supabase account (for vector DB).
• Deployment: Vercel deployment is supported via a one-click button, requiring specific environment variable configuration (NEXT_PUBLIC_CHAT_FILES_UPLOAD_PATH=/tmp).

Highlighted Details

• Supports chat with GPT-3.5.
• Leverages LangchainJS and Supabase vector DB for file-based conversations.
• Vercel deployment option available.

Maintenance & Community

The project is sponsored by Zhang Andy and features a star history link. Further sponsorship methods are available on the author's website.

Licensing & Compatibility

The repository does not explicitly state a license in the provided README.

Limitations & Caveats

The README does not specify any limitations or known issues. The project appears to be in active development, with features like "Chat with GPT-3.5" marked as complete.