Feature request: Support for custom CJK fonts (Korean) in drawtext

[woolfie1101]

Hi Rendi team,

First of all, thanks a lot for building such a clean FFmpeg API and for the excellent cheatsheet – it’s been really helpful for automating my video workflows.

I’m currently using Rendi to:

concatenate multiple short-form videos (with audio), and

add a persistent title banner at the top using the drawbox + drawtext pattern from the cheatsheet.
​

For English text this works great with the default fonts.
However, for Korean (Hangul) text, I run into the limitation that there is no straightforward way to use a CJK-capable font with drawtext in the hosted environment.

What I’d love to have:

Built-in Korean (CJK) fonts available in the container

e.g. something like fontfile=/usr/share/fonts/NotoSansKR-Regular.ttf

or a documented list of pre-installed fonts that support Korean.

Or, a documented way to use a custom font file from input_files with drawtext

For example, if I pass a TTF via input_files (e.g. in_font), a recommended pattern to

download/place it in a known path, and

reference it reliably from drawtext’s fontfile= in ffmpeg_command.

My use case is typical “shorts/Reels/TikTok” style videos where I want a bar like:

(Korean) [yellow]줄리엣[/yellow]의 AI 건축

showing at the top for the full duration, with mixed Korean + ASCII text.

Right now I can get the video logic working (concat + trim + overlays) but I’m blocked on proper Korean rendering because the sample fonts don’t contain Hangul glyphs. Having either:

at least one pre-installed Korean-capable font, or

an official example of how to reference a custom TTF/OTF in fontfile= when using input_files

would completely unblock this use case and make Rendi very attractive for Korean and other CJK creators.

If you need a minimal reproducible example or a sample TTF (e.g. the one I’m using from Supabase Storage), I’m happy to share it.

Thanks again for the great product, and I’d really appreciate CJK font support on your roadmap!

[peternaf]

Thanks for the feedback @woolfie1101 !

To add custom fonts you can use our dynamic input files feature:
https://docs.rendi.dev/dynamic-files#dynamic-inputs - this allows you to send us a folder as a zip file and within the folder you can add your fonts and other media files as needed.

We will add your specific example to the documentation.

[woolfie1101]

Hi, I’m trying to use the dynamic input files feature to add a custom Korean font to my video, but I keep getting a 422 error.

I’m calling the run-ffmpeg-command endpoint with this body:

{
  "input_files": {
    "in_1": "https://tempfile.aiquickdraw.com/v/240cad0cbce8d59addd908e4b287ed3e_1771233965.mp4"
  },
  "input_compressed_folder": "https://nuejpkpytqdhibwjdymt.supabase.co/storage/v1/object/public/n8n/n8n/Hakgyoansim%20Dunggeunmiso%20TTF%20R.zip",
  "output_files": {
    "out_1": "output.mp4"
  },
  "ffmpeg_command": "-i {{in_1}} -vf \"pad=iw:ih+400:0:200:color=black,drawtext=fontfile='Hakgyoansim Dunggeunmiso TTF R.ttf':text='라피스 라줄리 천연원석':fontcolor=white:fontsize=60:x=(w-text_w)/2:y=60,drawtext=fontfile='Hakgyoansim Dunggeunmiso TTF R.ttf':text='AI 줄리엣':fontcolor=white:fontsize=60:x=(w-text_w)/2:y=h-80\" -c:v libx264 -pix_fmt yuv420p {{out_1}}",
  "vcpu_count": 4
}

And I get this error:

422 - Exactly one of input_files or input_compressed_folder must be provided

My goal is:

• Input video: in_1 (vertical video, 1:1 or 9:16)
• Add black bars at the top/bottom
• Draw Korean text using the custom font inside the zip:

1. Top text: 라피스 라줄리 천연원석
2. Bottom text: AI 줄리엣

Could you please tell me:

1. Is it possible to use both input_files and input_compressed_folder together (video from input_files, font from the zip)?
2. If yes, what is the correct JSON structure and ffmpeg_command to access a font file inside the zip (e.g. Hakgyoansim Dunggeunmiso TTF R.ttf)?
3. If not, what is the recommended way to add a custom Korean font using the dynamic input files feature so that I can draw text on top of the video?

Thank you!

I’m also using n8n, so I would appreciate an example that works well with the run-ffmpeg-command endpoint from a workflow tool.

[peternaf]

Hi @woolfie1101

When using dynamic input files - all your input files should be in the compressed folder, don't use the "input_files" field
Meaning, both your media file and font file should be in the zip you are sending rendi

Answers to your questions

1. No
2. Use the "input_compressed_folder" field like you did with a zip that has two files
3. See 2.

We have this youtube rendi + n8n ffmpeg tutorial - https://youtu.be/Zikf4lLHWO0?si=71xkRsPDcubclho1
What would you specifically like to see there which is missing?

[woolfie1101]

Hi, just wanted to say thank you –
after following your explanation about putting both the video and the font TTF into the same zip and using only input_compressed_folder, I got it working perfectly.

I can now render Korean text on top of the video using my custom font with Rendi + n8n.
Really appreciate the quick and clear support! 🥹❤️

[peternaf]

Happy all worked out and you are able to get your automation going 🎉