LinkedIn Slop Detector

This repo is for the Codédex monthly challenge. The full tutorial can be found here: TUTORIAL.md.

Note

If you're here for the tutorial, you'll need to download two files from this repo and drop them into your project folder: the emoji font NotoColorEmoji.ttf (so your card looks the same on every computer) and the card generator card.py.

It's a Python tool that reads any LinkedIn post and gives it a Slop Score /100 with a verdict, then saves it as a shareable card. Along the way you'll learn to use the Hugging Face API for zero-shot text classification and blend AI judgment with your own transparent rules.

Quick start

python3 -m venv .venv
source .venv/bin/activate
pip install requests python-dotenv Pillow

Add your Hugging Face token to a .env file (see the tutorial for how to get one):

HF_TOKEN=hf_your_token_here

Then run it:

python slop.py

What's in here

File	What it does
slop.py	The main detector: rule signals + Hugging Face scoring
card.py	Draws the shareable score card with Pillow
NotoColorEmoji.ttf	Emoji font so the card looks the same on every computer
TUTORIAL.md	The full step-by-step build tutorial
WRITE-UP.md	Behind the scenes: how the idea came together and the challenges I hit
images/	Screenshots and example cards used in the tutorial

More resources

• Hugging Face Inference API docs
• Zero-shot classification explained
• Pillow documentation