Use Vidrank to prioritize which YouTube videos to watch first. Vidrank presents you with small batches of videos. You can classify a subset of each batch as better than the rest and your selections are processed by the TrueSkill algorithm to produce ratings. Ratings are used to determine the next best set of videos to consider to maximize information gain and decrease rating uncertainty.
If you have a backlog of videos waiting in your Watch Later playlist, then this tool was made for you.
Requirements:
Install the backend package
cd backend
poetry install
Install the frontend package
cd frontend
pnpm install
Set a few environment variables that are used by the server.
You will need an API key for the YouTube Data API (v3)
export YOUTUBE_API_KEY="<youtube-api-key>"
export VIDRANK_CACHE_DIR="<path-to-cache-directory>"
export VIDRANK_PLAYLIST_ID="<youtube-playlist-id>"
Start the backend
fastapi dev backend/vidrank/app/app.py --reload
Start the frontend
pnpm run dev
LOG_LEVEL=DEBUG fastapi dev backend/vidrank/app/app.py --reload