Spectacle 6.5.5 on Fedora 43 missing OCR “Extract Text” feature despite Tesseract being installed

I couldn’t configure the spectacle in my fedora 43 plasma as extract text option is enabled.

version of the apps:

(base) omerg@fedora:~$ tesseract --version
tesseract 5.5.2
 leptonica-1.87.0
  libgif 5.2.2 : libjpeg 6b (libjpeg-turbo 3.1.2) : libpng 1.6.53 : libtiff 4.7.1 : zlib 1.3.1.zlib-ng : libwebp 1.6.0
 Found AVX2
 Found AVX
 Found FMA
 Found SSE4.1
 Found libcurl/8.15.0 OpenSSL/3.5.4 zlib/1.3.1.zlib-ng libidn2/2.3.8 nghttp2/1.66.0
(base) omerg@fedora:~$ spectacle --version
spectacle 6.5.5
QThreadStorage: entry 2 destroyed before end of thread 0x55811a644480
QThreadStorage: entry 1 destroyed before end of thread 0x55811a644480
(base) omerg@fedora:~$ 

what am I missing? can anyone help me?

That feature is coming in Plasma 6.6, so it’s not in Fedora 43 yet.

If you’re confident installing a beta of Plasma 6.6 from a COPR, you could test it that way:

I tested this in Rawhide (which also has the Plasma 6.6 beta) and unfortunately I can’t get it working.

The Spectacle options screens show that OCR “should” be available, but it fails to find the necessary tesseract libs even though the package is installed.

Bug ticket:

I just installed the latest packages as of today. Upgraded to Plasma 6.6, Fedora 43. No OCR in Spectacle.

spectacle.x86_64

1:6.6.0-1.fc43

I had the same issue, updated today. Got a prompt with all the new KDE features including OCR. But it was missing from spectacle.
Fedora 43 KDE 6.6.0

Fixed it by installing:

sudo dnf install tesseract-devel
sudo dnf install tesseract

It’s possbile to install language packs, find them by using:

sudo dnf search tesseract-langpack

Nice find!

tesseract-devel seems to install /usr/lib64/libtesseract.so, which Spectacle can find (unlike the library files from plain tesseract).

If you run spectacle on the command line, you’ll see exactly which library it’s looking for; in my case (Fedora 43, updated wih KDE 6.6 in 2026-02-18), it’s libtesseract.so.5 .

The tesseract-libs-5.5.2-1.fc43.x86_64.rpm package, however, constains only libtesseract.so.5.5 and libtesseract.so.5.5.2, both in /usr/lib64 .

So, you can fix this issue with “cd /usr/lib64 ; sudo ln -s libtesseract.so.5.5 libtesseract.so.5” .

Of course this is an ugly hack, but for me it’s better than installing several *-devel packages just to get that one link, which probabbly should be in the tesseract-libs package to start with.

the workaround published today seems to work well. At least it’s ok on my fedora 43

With the updated spectacle-6.6.0-2.fc43, OCR should now work as long as tesseract is installed, with no need for tesseract-devel.

I still needed to install “tesseract-devel” using “spectacle-1:6.6.3-1.fc43”.

Good catch, this looks like a regression.

I can reproduce - I have the following packages installed (note, no tesseract-devel):

$ dnf list --installed "tesseract*"
Installed packages
tesseract-common.noarch       5.5.2-1.fc43  updates
tesseract-langpack-eng.noarch 4.1.0-11.fc43 fedora
tesseract-libs.x86_64         5.5.2-1.fc43  updates
tesseract-tessdata-doc.noarch 4.1.0-11.fc43 fedora

Previously (as of 6.6.0-2) this was enough for Spectacle to provide OCR. With 6.6.3-1 it doesn’t work and tells me “OCR functionality is not available”.

(tesseract-libs should be sufficient to provide the necessary libraries, but I did try installing tesseract and that didn’t fix it either.)

Reported this on the upstream KDE bug, and reopened the RHBZ bug.

Funnily enough, I noticed this last night when I clocked that Spectacle was no longer offering OCR, when it had done last time I looked. I assumed I’d inadvertently removed the tesseract-devel package.

Nice to know the dementia isn’t kicking in and it really is a regression.

I can reproduce as well.

Fixed on 6.6.3-2