Getting Kdenlive & Whisper Speech to Text to work, 7/2025

I had a similar issue as seen here:

Can’t install whisper using kdenlive’s interface. Unable to create python virtual environment (Kdenlive Whisper OpenAI speech to text )

…and I wanted to provide a potential solution in case the solution there doesn’t work for you, like it didn’t for me.

You might see, after reading this, that nothing is wrong with the above solution, just that things might’ve changed over the past few months between KdenLive and OpenSUSE (it seems more Kdenlive).

Note: && TLDR: Kdenlive might require a different Python version in the future. The key is that you can install multiple versions side‑by‑side on openSUSE (python311, python312, python313, etc.) so Kdenlive can find the one it needs.

I tried the solution at the above link, which was to install the python virtual-env package through Zipper/YAST, but I continued to see “cannot find system Python” in the Kdenlive window where you add models for speech to text.

What I ended up doing was running Kdenlive from the commandline (by entering kdenlive in the prompt), so that I could see what was happening as I clicked on buttons in the application. When I pressed the button that kept yielding “cannot find system Python” in Kdenlive, the terminal yielded the following:

Cannot create the python virtual environment:

pyenv: python3.12: command not found

The python3.12' command exists in these Python versions:
  3.12.8

Note: See 'pyenv help global' for tips on allowing both
      python2 and python3 to be found.


Kdenlive is looking for Python 3.12.8. I haven’t updated my install of Tumbleweed in at least a week, but the system version is at 3.13.5. Because of Tumbleweed’s rolling nature, the system’s python is ahead of the specific version of Python Kdenlive requires.

Aside:
You may ask, “You have pyenv installed, why not just use pyenv to set a global version?”

And my answer is while I have pyenv installed mostly for setting versions of python locally in directories for AI and other experiments, I was hesitant to set a global version of python through pyenv because then I’d be on the hook to update the global version as newer versions of Kdenlive released.

But back to the main topic

If you’re new to openSUSE, or Python, and how the two are made to work together, when it comes to baremetal python that doesn’t run in a virtual environment, it seems you can install multiple versions of Python on the system via YaST or Zypper, for situations just like this. So I looked in yast and added python312.

If YaST isn’t your thing, you can also enter
sudo zipper install python312 or whatever version of python may mismatch with Kdenlive in the future.

The version that installed was 3.12.11, not 3.12.8, but I figured it was worth a try. It turned out that was sufficient, and I was then able to add whisper models for speech to text in Kdenlive.

I hope this helps!

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.