GPicTionary - Pictionary game assisted by GPT

This game was quickly POCed and developed for an AllHands presentation at Woflow. By quickly POCed I mean it was just eXtreme Go-Horse process

Install Nodejs dependencies

npm install

Copy .env.example to .env and fill in the required environment variables

cp .env.example .env

You will need to fill the following environment variables, which you can get by either running supabase locally or by going to your project settings in the Supabase dashboard:

NEXT_PUBLIC_SUPABASE_URL=
NEXT_PUBLIC_SUPABASE_ANON_KEY=

Getting Started with NextJS server

First, run the development server:

npm run dev

Project Structure

Credits

Project was bootstrapped from supabase realtime example:

Learnings

GPT can be nice, but specific purpose models can be better, for example: https://github.com/ximinng/SVGDreamer
Streaming data with supabase: https://www.youtube.com/watch?v=9N66JBRLNYU
Nextjs App Router: https://nextjs.org/docs/app/building-your-application/routing
Embeddings 101: https://youtu.be/Yhtjd7yGGGA?t=821
PGVector: https://github.com/pgvector/pgvector