papyrus

a simple RSS feed aggregator that learns what you like to read

RSS feeds are great, but they can also be a firehose. What if we could recommend articles based on what you like?

Papyrus uses a super-simple recommendation system (SVM over TF-IDF features) to learn your preferences.

Setup

• Requires python3 and node.js
• Pick a port for the frontend and backend
  • Default: 2400 for frontend, 2430 for backend
• Configure .env, server/Makefile, and start.sh accordingly
  • I will probably make this easier in the future

# .env
VITE_FRONTEND_PORT=<YOUR_FRONTEND_PORT>
VITE_BACKEND_URL=http://localhost:<YOUR_BACKEND_PORT>/api

# server/Makefile
dev:
	uvicorn app:app --reload --host 0.0.0.0 --port <YOUR_BACKEND_PORT>

prod:
	uvicorn app:app --host 0.0.0.0 --port <YOUR_BACKEND_PORT>

# start.sh
yarn preview --port <YOUR_FRONTEND_PORT> &

Then run ./start.sh in the background (either tmux or append &)

The database lives in server/data, in case you want to back it up. I use an NGINX reverse proxy to serve the frontend/backend + add password protection.

Disclaimers/Notes

1. This was hacked together in my spare time over ~1 week
2. A lot of this code is underoptimized; I'm working on making page requests faster
3. A lot of this code is LLM generated. I want to see how far I can push current tools.
4. The style is the same as my blog - derived from bearblog and flexoki