Shotpad 1.0.7

Shotpad can now read text out of a screenshot. Pick Grab text (G),
drag a box over some words, and they are on your clipboard — an error code, a
licence key, a URL that only exists as pixels.

Added

• Grab text, the thirteenth tool. Straight to the clipboard: you chose the
  tool and swept a region, so the copy is the thing you asked for. The status
  bar reports what was taken and says so plainly when the reading is doubtful,
  because a silently wrong answer is the failure that costs you.
• English, Turkish, Azerbaijani and Russian. The picker follows your system
  language rather than defaulting to English. More can be dropped into
  ~/.local/share/shotpad/tessdata without waiting for a release.
• Text you have redacted stays hidden. Grab text reads the picture as it
  will be exported, so a blur or a black box is not quietly undone by the tool
  sitting next to it in the rail.
• Recognising nothing leaves your clipboard untouched — wiping it with an
  empty result is the worst outcome of a stray drag.

Notes

The recogniser is a single static binary with no dynamic dependencies, built
without the network and archive backends a screenshot tool never reaches. That
keeps it off the OpenSSL and Kerberos treadmill those would have brought, and
means it has no glibc floor at all: it runs wherever the rest of the AppImage
does. The download grows by about 10 MB, most of it language models.

Small text is enlarged before reading, by an amount measured from the image
rather than guessed — without it, 1× interface text reads "Comer radius" for
"Corner radius".

The interface itself is unchanged and still English; the languages above are
only about what can be read out of a picture.

Updating

pipx upgrade shotpad

AppImage users: download below, then re-run --install so the menu entry
points at the new file.

chmod +x Shotpad-1.0.7-x86_64.AppImage
./Shotpad-1.0.7-x86_64.AppImage --install