EthicalEye is an open-source content moderation model I created, built on XLM-RoBERTa and supporting 88 languages. It classifies harmful and unsafe content across languages with a single model, making moderation practical for products with international audiences.
Released under the Apache 2.0 license, EthicalEye has passed 10,000 downloads on Hugging Face, was featured in the Hugging Face community showcase, and has been adopted by companies for production content moderation, reducing manual review workloads by approximately 70%.