The rise in social media users in recent years has made community content moderation difficult. Here's my attempt at changing that, with an open-source Discord bot that aims to help moderators keep their communities safer.

Puffin is an auto-moderation Discord bot utilizing sentiment analysis and zero-shot learning. Please allow a few seconds for a message to be flagged in a text channel. Feel free to reach out with feature requests, bug reports, or any other comments/issues you may have.

# Tech Stack:
- Puffin is written using `discord.py`
- Database created with `MongoDB` - all data is encrypted in the cloud
- Hosted on a Raspberry Pi 5
- ML models integrated using Hugging Face Pipelines

# Methodology:
Puffin filters messages based on whether or not they hold specific emotions, their sentiment, zero-shot labels that the user specifies, and more. Flagged messages are added to a moderation queue that can be viewed via an ephemeral message if you have the right permissions.

# GitHub Repo:
[https://github.com/Windshield-Viper/Puffin-Discord-Bot](https://github.com/Windshield-Viper/Puffin-Discord-Bot)

Puffin

AI-based automoderation - natural language processing, zero-shot learning, Toxicity API, and more for safety. Open-source and customizable. 

Puffin 투표

저희를 지원하고 싶으십니까?