Plate A — shapes, bars, figures
Plate B — dense text
Live DOM text — exercises the tab-capture path
This block is real DOM, not an image, so boxing it forces the capture path through
getDisplayMedia. Chrome grants the current tab without a picker when
preferCurrentTab is set; the stream is kept alive afterwards so the second
capture is instant.
The extractor should resolve these as text lines rather than shapes: similar heights, shared baselines, tight horizontal gaps. If line count comes back near zero, the crop is probably too small for the 512-pixel analysis raster.
Boxing the swatch strip below should instead produce a handful of wide rectangles and a six-entry palette with near-equal shares.
Last descriptor
Nothing captured yet. Drag the reticle over a plate, or paste an image.