[duckb]

Does this support table detection and extraction?

[fzysingularity]

Yes, it's experimental at the moment: https://docs.vlm.run/guides/doc-ai/guide-visual-grounding