sweic / filegpt

Chat with your PDF/MP3 files using ChatGPT and Embeddings

Live Web App

📝 Project Description

Upload your PDF/MP3 files and chat with its contents. The data is converted into embeddings and stored in a vector database, Pinecone. Queries are made with the help of Langchain and OpenAI's embeddings search.

✅ Features

• Familiar ChatGPT UI
• Upload and stores files (PDF/MP3 supported) into Pinecone
• Make conversation and queries to OpenAI's GPT3.5 with knowledge on the file's content
• Manage conversations and files
• Persistent conversations and files using Local Storage

📚 Development Stack:

• SolidJS
• ExpressJS
• NodeJS
• Pinecone

🛠 Setting up local development environment:

• Create a Pinecone account, and setup index
• Create a new folder and git clone https://github.com/sweic/filegpt.git
• Fill up the .env.example file inside client and server subfolder and rename to .env
• npm install in the client subfolder
• To start client, run npm run dev in client subfolder
• App should be running on http://localhost:3000/
• To start server, run npm run nodemon

🛣️ Roadmap

• Better error handling