Comment Moderation Classifier

Clean your comment section automatically

Classify comments as SPAM, HATE, CONSTRUCTIVE, PRAISE, or OFF_TOPIC. Keep the good, lose the noise.

Primary use cases

YouTube/TikTok moderation
Community health tracking
Sentiment reporting

Run this on your prompt →Try the classifier live

Why you can’t just prompt-engineer this yourself

The evolution story

Gatekeeper was bred against the hardest moderation boundary: sarcastic praise vs. constructive criticism, and SPAM vs OFF_TOPIC. Handles short-form platform tone (TikTok, Instagram) natively.

Live demo

Try Gatekeeper — Gen 1.0

0/8000 characters

Free trial: 10 runs across 7 days · No credit card · No signup needed to start

Evolutionary lineage

Gatekeeper is a Gen 1.0 agent

1.0

Generation 1.02026-04-14

Initial evolved release. 87% holdout accuracy.

Measured accuracy: 87%

This is one of 11 case studies proving the Cambrian evolution engine works. Your prompt can follow the same arc.

Want these numbers on your own prompt?

Bring your production prompt and a labeled dataset. We’ll evolve it the same way we evolved Gatekeeper. First evolution is free.

Evolve my prompt →