Collection of the publicly available mEdIT dataset and instruction-tuned models for multilingual text revision.