About

This is a fullstack Q&A app that answers question based on the text documents provided to it. It uses LangChain to communicate with OpenAI API to interpret questions and give answers.

Stack

Requirements

1. Auth0
   Create a Single Page Application in Auth0.com > grab application's domain and cleint ID.
2. MongoDB
   Create a MongoDB Atlas project > grab connection string and admin password.
3. OpenAI
   Create an OpenAI account > grab your API key.
4. in the server directory create a .env file and set these values:

    ISSUER = {AUTH0-APPLICATION-DOMAIN}  
    OPENAI_API_KEY = {YOUR-OPENAI-API-KEY}  
    MONGOPASSWORD = {YOUR-MONGODB-PASSWORD}  

5. in the client directory create a .env file and set these values

    VITE_DOMAIN = {AUTH0-APPLICATION-DOMAIN}  
    VITE_CLIENT_ID = {AUTH0-APPLICATION-CLIENT-ID}  

Starting Up

cd server
npm start

then, in a new terminal,

cd client
npm run dev

Documents

The app currently reads from two provided sample documents about Spinosaurus (of course). Namely, "about.txt" and "description.txt". You can replace your own documents and change the corresponding paths in server/src/AI/QA.ts, so that it gives answers according to your documents.