Fast, ultra-accurate text extraction from any image or PDF—including challenging ones—with structured markdown output powered by vision models.