unitaryai/detoxify

Trained models & code to predict toxic comments on all 3 Jigsaw Toxic Comment Challenges. Built using ⚡ Pytorch Lightning and 🤗 Transformers. For access to our API, please email us at contact@unitary.ai.

69
/ 100
Established

This project helps content moderators, community managers, and platform administrators automatically identify harmful content in online discussions. You input a comment or a list of comments, and it tells you if each comment is toxic, severely toxic, obscene, a threat, an insult, an identity attack, or sexually explicit. It's designed for anyone managing online communities or user-generated content.

1,202 stars. Used by 4 other packages. Actively maintained with 2 commits in the last 30 days. Available on PyPI.

Use this if you need to quickly flag potentially harmful comments across multiple languages to maintain a safe and respectful online environment.

Not ideal if you need to perfectly capture nuanced humor or self-deprecating comments, as it may misclassify them as toxic.

content-moderation community-management online-safety social-media-management brand-reputation
Maintenance 9 / 25
Adoption 14 / 25
Maturity 25 / 25
Community 21 / 25

How are scores calculated?

Stars

1,202

Forks

141

Language

Python

License

Apache-2.0

Last pushed

Jan 05, 2026

Commits (30d)

2

Dependencies

3

Reverse dependents

4

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/nlp/unitaryai/detoxify"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.