Has anyone tried using Grok to read text from uploaded screenshots, signs, or product packaging? I’m interested in ordinary, practical OCR—things like pulling a short serial number from a photo or reading a label that isn’t perfectly straight—not counting objects or interpreting a whole scene.

How does it cope with small print, glare, and cropped text? I’m trying to gauge when text in screenshots is clear enough to trust and when it’s better to run a dedicated OCR tool.