Region capture is drag-to-select: draw a box over the table, chart, or error message you care about and get exactly that image, with text kept large enough for a model to read.
Region-capture the table, run the extract-as-table action, and paste clean CSV or Markdown straight into a spreadsheet or document. Headers, rows, and columns come back structured.
Yes. Captures are real screenshots of rendered pixels, so canvas dashboards, charts, images, and in-page PDFs are captured exactly as displayed, even where the DOM is empty.
Full-page capture stops at 60 frames, roughly 60 viewport heights, and tells you when it truncates. For longer pages, capture the specific sections you need with region capture.