dots.mocr is a focused, no-nonsense specialist built for optical character recognition tasks, handling both text and image inputs to extract and interpret written content from visual sources. It brings a long context window of 131K tokens to the table, meaning it can process substantial documents or image batches without losing track.