Error when creating Korean PDF

When you select Pdf Generator in Connected Actions and create a PDF, there is an error where Korean text is broken. However, English letters and numbers appear correctly.

If you’re using the generic Generate PDF action, try cloning it and customizing it to your use-case > Try to add a Korean language package to the “Additional packages” section under “Action Details” (See screenshot).

[Korean language Python package](KoNLPy: Korean NLP in Python — KoNLPy 0.6.0 documentation

In this case, I believe you want to type in “KoNLPy”, and then hit the “+”.

Then you can go to the “Action code” section and use Python code to import it. Use ChatGPT 4o to generate the code.

Also, try changing the XML Pdf package over to “fpdf”. I found it generates more stable PDFs over multiple uses.

Check out this relevant thread for more info.

1 Like

hi my friend. did you find the way to fix it in korean language? i am trying to find out the same for greek language. i will be grateful for sharing the solution

1 Like

Same here with any non-latin letters.

1 Like

@admin_mike any updates?

We dont have any updates on rendering korean characters in PDF right now. If you configure the action yourself, there are probably some tricks you could do. The PDF action uses HTML to render korean characters. I would suggest talking with ChatGPT and seeing if there are any tricks you could use to achieve this.

In my case, it’s not Korean, it’s Cyrillic letters. So I think the problem is much broader than just a particular language. It could be literally everything that’s outside Latin characters.

@admin_mike what library do you use as a tool for generating PDF? Maybe I can look into this issue myself and come up with solution.
UPD: I figured it out; it’s xhtml2pdf library.

Again: anything except Latin characters is rendering as black rectangles. This is important; lots of people potentially affected.

UPD: I have great news! The issue is that the font used in XHTML by default does not support multilingual characters.

There are a lot of discussions on non-Latin characters when using xhtml2pdf library. And everyone is telling that one should simply include a multilingual font in the HTML file. That’s it!

Noto Sans seems promising. https://fonts.google.com/noto/specimen/Noto+Sans

Hi @conrad,

Try sending a prompt like the following:

Create me a pdf about AI powered tools with this font https://fonts.googleapis.com/css2?family=Noto+Sans:ital,wght@0,100..900;1,100..900&family=Pacifico&display=swap
1 Like

IT WORKED! Thank you!
For others: I added instructions and a link to the font into the “Functions input” description field.

1 Like