A unified text-to-animation model that generates articulated motion for any skeleton, with training, inference, dataset pipelines and editing tools.