This Week in Plasma: OCR in Spectacle and many UI improvements - KDE Blogs

Welcome to a new issue of This Week in Plasma!

This week Spectacle gained OCR (optical character recognition) functionality, allowing you to turn words in images into selectable text!


This is a companion discussion topic for the original entry at https://blogs.kde.org/2025/11/15/this-week-in-plasma-ocr-in-spectacle-and-many-ui-improvements
10 Likes

Ooh, if the OCR works really well, then that would help as I no longer need an external ones. Does it use tesseract, and if so, what models and languages are used? Can user change them? I often use OCR on English, Indonesian, and Japanese (horizontal and vertical) languages you see.

Renaming a file or folder on the desktop now lets it keep its existing position. (Błażej Szczygieł, link)

This is also very helpful! An annoyance on Windows but also on Linux as well when I’m renaming a lot of files. Thanks!

When an app asks to register keyboard shortcuts on launch and you don’t let it, this preference is now remembered, instead of the annoying app just asking again the next time it launches.

Hm, what if someone accidentally declines or just hurriedly click decline because they just want to do something else? I think a “Don’t ask again” toggle would help - I for one think it’s better to add that one extra click for permanent permissions, it’s an annoyance and people might just eventually auto-pilot it, but the extra one click might get more people to think from time to time.

Another thing that could help is a permission request log of some sort. Is there already such a thing implemented or being worked on?

But overall a really great list of improvements. Thank you for everyone’s hardwork!

I know this is not a feature request page, but, I would love to see inside the new device notification the mounting point, or the serial device inserted (e.g. /dev/USB0) or screen 1. I know it can be done easily, as I had that in my own implementation before this was implemented.

Regards,

Loren

It is FOSS, you can check by yourself and tell the world: Add OCR functionality to Spectacle (!462) · Merge requests · Plasma / Spectacle · GitLab

Allowing users to extract text from screenshots using Tesseract OCR engine.

The video shows how one can change the language used.

Although the available languages depend on your tesseract installation.

2 Likes

Nate, are there any planned updates or planned feature enhancements to the Plasma Keyboard? I just started using it instead of Maliit on my Surface Pro tablet PC. Just like with Maliit, Plasma Keyboard is missing features needed to support Plasma and Linux desktop computing, like function keys, special keys (Alt, Tab, etc), and arrow keys. I’m willing to test new features for Plasma Keyboard and provide feedback, if needed.

Plasma Keyboard is, to my knowledge, a fairly actively developed project. I recall that emulation of physical modifier and other special keys is planned in some capacity. But I’m not active in the project, so all I have is that fuzzy impression, sorry!

Maybe offtopic, but is KDE planning to change anything in Plasma for the Steam Frame’s desktop mode? Does KDE (or Techpaladin) have access to the upcoming Steam Hardware currently?

I’m sorry to raise this old thread, but I was looking at spectacle ocr and I was wondering why would not have it on Spactacle 6.6.2? Im on CachyOS, KDE Plasma 6.6.2, X11 (the option is not even there). Is this maybe a Wayland only feature?

I suspect the same happens with other fixes described in the blog, like “drag a tab out of a Chromium/Chrome window and immediately tile it to a screen edge or corner. (David Redondo, link)” that neitehr are available on X11.

You’re probably missing the tesseract package.

Indeed, that fixed the Spectacle not having OCR! Thank you very much :clap:
Some sort of user indication somewhere would be nice. If not for that blog post, I would never know OCR was supported by the app.

Are there any plans to process QRs and bar codes?

As or the the chromium-based-browser tab-window tiling, was that fixed only for Wayland?

Things like that are usually handled by the package manager. On Arch-based distros like yours, whilst installing updates, pacman will show messages for new optional dependencies. Also, the package information will show dependencies:

$ pacman -Qi spectacle 
Name            : spectacle
Version         : 1:6.6.2-2
Description     : KDE screenshot capture utility
Architecture    : x86_64
URL             : https://kde.org/plasma-desktop/
Licenses        : GPL-2.0-or-later
Groups          : plasma
Provides        : None
Depends On      : gcc-libs  glibc  kconfig  kconfigwidgets  kcoreaddons  kcrash  kdbusaddons  kglobalaccel  kguiaddons  ki18n  kio
                  kirigami  kjobwidgets  knotifications  kpipewire  kquickimageeditor  kservice  kstatusnotifieritem  kwidgetsaddons
                  kwindowsystem  kxmlgui  layer-shell-qt  libxcb  opencv  prison  purpose  qt6-base  qt6-declarative  qt6-imageformats
                  qt6-multimedia  wayland  xcb-util  xcb-util-cursor  xcb-util-image
Optional Deps   : tesseract: OCR text extraction [installed]
Required By     : plasma-meta
Optional For    : None
Conflicts With  : None
Replaces        : None
Installed Size  : 6.03 MiB
Packager        : Antonio Rojas <arojas@archlinux.org>
Build Date      : Do 05 Mär 2026 19:28:28
Install Date    : Sa 07 Mär 2026 18:47:51
Install Reason  : Installed as a dependency for another package
Install Script  : No
Validated By    : Signature

Note the “Optional Deps” line.

1 Like

I don’t have an install to test with now that reproduces the problem. But back when Spectacle had trouble finding Fedora-packaged Tesseract:

  • In Spectacle’s Options >> Configure Spectacle dialog, a message appeared stating that OCR wasn’t available
  • Next to that message was an (i) button which, when clicked, popped up a message:

Please install the required packages: * tesseract * tesseract language data (e.g., tesseract-ocr-eng for English).

1 Like

I think the blog post should be updated to mention Tesseract, ideally somewhere near the top part mentioning Spectacle OCR support. Back when the blog post was released I was so confused why I don’t see the extract text option, and assumed it’s still being rollout or something.

1 Like

It’s already got that too! You need the prison framework installed.

1 Like

Oh lordy, this is fantastic!
A bit confusing to be honest, but fantastic, thank you! :folded_hands:

Seems like the QR reading has an entirely different ux approach than OCR, after installing prison package, I had to test a few times until grasping how it works.

The marketing material falling a bit short imho :sweat_smile:
Spectacle has a fantastic feature set, I dont know why these are so buried in way you’d never find out.

You’re welcome!

If your distro packages Spectacle well, they aren’t very hidden. :slight_smile:

You could say it’s our fault for making these dependencies optional, which allows for distros to package spectacle in a way that leaves it up to users to ensure they have all the optional dependencies.

That may be acceptable for users of Arch-based distros who are expected to be actually reading pacman output and noticing uninstalled optional dependencies, but it’s less acceptable for more user-friendly distros like Kubuntu or Fedora; I would expect these distros to be making the optional dependencies mandatory.

However since you’re on CachyOS, you’re using an Arch-based distro so I’m afraid the responsibility is on you. :slight_smile:

Yes I agree, though thats not what I meant. There is no single point of info, easy to find site, or “promo material” lets call it, that describes the Spectacle feature set. Is a sort of self discovery process, that feels like a pitty since it has so much to offer.

Anyway, Im a happy user now :slight_smile:

Yeah, that’s a general issue with KDE software: it evolves faster than its documentation.