Why would I capture a whole screen instead of one chart?

Because the reason you took the trade is rarely all in one chart. A full-screen shot catches the higher timeframe, the order-flow pane and the depth ladder that were also on your desk — and the AI chart read is built to use every one of them.

Capture photographs a display, not a window. That looks like a limitation and is closer to the opposite.

Think about what is actually on your screen at the moment of a fill. A 1-minute chart you entered on. A 30-minute chart that told you where the level was. A cumulative-delta pane. Maybe a second instrument you were watching for confirmation. You did not decide from one of those; you decided from the arrangement.

A single-window capture keeps the 1-minute chart and throws the rest away. Three weeks later the still shows an entry that looks unmotivated, because the thing that motivated it was in a panel that is not in the picture.

The AI read is explicitly built for the wide shot. When a chart image is attached to a trade, the vision pass is told the capture may be a full-screen shot of a multi-chart template, and to treat every panel as evidence. It identifies each panel by symbol and timeframe, reports back which ones it found, and reads levels from all of them. Its instructions are unambiguous that no timeframe outranks another: a 30-minute moving average and a 1-minute point of control are equally valid reasons to have taken a trade.

Two limits worth knowing. Panels showing a different instrument are ignored — a different symbol on screen is context for you, not evidence for this trade. And a level that appears on three timeframes is reported once, not three times.