Sounds easy, ain't it?
As a matter of fact it wasn't, but it works.

What you need:

• Standard Language File

• OpenAI as paid service

• DeepL as paid service

How to do it:

The Standard Language File is an xml File with word tags. Within the word tags, there are phrases, which can be empty, have standard characters as text or HTML-code.
OpenAI, Gemini and Grok tell you, that they can give you script written in python, because they don't want to translate a file that long. If you try to argue them to translate the file, all AI will tell you a whole lot of nonsense, give you only parts of what you want and give you a pack of lies and excuses.
The scripts they write won't work, because they will crash the structure of the XML, mainly the CDATA and HTML tags will be broken.

After fiddeling with some examples it looks like you can give OpenAI the standard language xml and tell it to extract all phrases in the word tags - even the empty ones and wrote each of it in a line of a new file. You now have a text file standard_language_extracted.txt. Next step is to open it in LibreOffice and save it as an Microsoft 365 docx. Now you upload the standard_language_extracted.docx to DeepL and translate it to the language you like, which gives you standard_language_extracted_lang.docx. We open this in LibreOffice too and save it as a standard txt file. Last step - at least we think it is - is to tell OpenAI to insert each line of the standard_language_extracted_lang.txt into the original file and save it as a new file.

You remember about the last step? It wasn't, because you have to double check, if OpenAI lost the first three and the last two lines of the original file and be smart and check a few lines with HTML, special characters and empty phrases and repair all broken parts of your new language file.

Now upload it and test it!

The two commands to OpenAI were:

Extract all the texts in the Word tags and write them line by line in a file. If the Word tag is empty, then create an empty line, so only a line break.

Put the lines in this document as text in the Word tags of the original file. Get the structure of the lines that contain CDATA and save everything in a new file.

Of course there are some inconsistencies in the translation. Some phrases will be translated correctly, but will not fit the setting it will be in and sound crude.

Nevertheless - the translation into German i did that way, works like a charme. Wanted to do something like that for many many years.

Better picture including reasoning: