Korean subtitle fonts for a video API: 29 faces on Sume
Sume's caption font field names 29 Hangul faces, all SIL OFL 1.1, from Pretendard to Gmarket Sans. Which suit which style, and what fails on a Latin style.

Sume's caption endpoint accepts 29 named Hangul fonts in its optional font field, from pretendard and do-hyeon to gmarket-sans, black-han-sans and nanum-pen. All are SIL Open Font License 1.1 and ship with the renderer, so there is nothing to upload or license separately for the listed names. Any other name is rejected rather than substituted.
The source for every claim here is Sume's video captions page. It describes each face in a few words, so the 'feel' column below repeats those descriptions and does not add our own opinion of design quality.
How does the font field interact with style and language?
style picks the look and the motion, design edits it, font picks the face it is drawn in, and language only tells speech-to-text what to expect. The docs state it plainly: language never selects the style or the font.
Omitting font keeps the style's own face: Pretendard for korean-ad, weight-shift, highlight, pill-karaoke and editorial-emphasis, and Do Hyeon for black-outline and clip-wipe. Setting font keeps the style and swaps the face. The field is for Hangul styles only; a Latin style such as slam with a Hangul font returns 400 with caption_font_requires_hangul_style.
Which faces are available?
The remaining faces are bagel-fat-one, dongle, yeon-sung, single-day, hi-melody, ibm-plex-sans-kr, gothic-a1, hahmlet, song-myung, poor-story, gamja-flower, stylish, sunflower, nanum-gothic, nanum-myeongjo, gaegu, cute-font and east-sea-dokdo, spanning display, script and workhorse categories.
| font value | Family | Documented feel |
|---|---|---|
| pretendard | Pretendard Variable | Clean baseline |
| do-hyeon | Do Hyeon | Thick rounded, the CapCut classic |
| black-han-sans | Black Han Sans | Impact |
| gmarket-sans | Gmarket Sans | Geometric retail display, the CapCut / YouTube title staple |
| gasoek-one | Gasoek One | Ultra-thick impact |
| jua | Jua | Soft cute rounded |
| dunggeunmo | DungGeunMo | Pixel / retro |
| nanum-pen | Nanum Pen Script | Handwritten |
| noto-sans-kr | Noto Sans KR | Workhorse sans |
| noto-serif-kr | Noto Serif KR | Workhorse serif |
| gowun-dodum | Gowun Dodum | Soft editorial |
What breaks the weight animation?
weight-shift and korean-ad animate the wght axis, and the docs say only Pretendard carries it. Put either style on a static face such as jua and the colour and scale emphasis stay but the weight travel is lost. If the weight change is the reason you chose the style, leave font unset.
editorial-emphasis is deliberately two-faced: its emphasis line is always Black Han Sans whatever font says, because the look is the contrast between the two faces. Setting font moves the lead line only.
How do you request a font?
Name a Hangul style, set font, and send Korean text or a Korean-speech video. The call below burns Gmarket Sans on the black-outline style. A style you name is rendered as named; Sume does not quietly switch to another one.
curl -X POST https://api.sume.com/v1/video-captions \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: ko-gmarket-001" \
-d '{
"video_url": "https://media.sume.com/artifacts/example/clean.mp4",
"style": "black-outline",
"font": "gmarket-sans",
"language": "ko"
}'Should you change the default face at all?
Often not. Each Hangul style was authored with its own face in mind, and the docs choose Pretendard for five of them and Do Hyeon for two. Changing font is worth it when a brand guideline names a face, or when you want a deliberately different mood: jua or hi-melody for soft, friendly copy, gasoek-one or black-han-sans for loud, short hooks, noto-serif-kr for an editorial or documentary tone. Those pairings are our reading of the documented feel column, not a tested recommendation.
A practical workflow is to render one caption job with the default face, then use source_caption_id with a different font and style to compare without paying for a second transcription. The restyle still bills as a render, so keep the experiment to two or three variants on a short clip. Long phrases in a heavy display face such as Gasoek One take more width, so check the design.phrasing limits (max_words, max_chars) and design.typography.safe_width_ratio if lines wrap more than you want.
Can you bring your own font file?
Not according to the docs. The list is closed: any other name is rejected, and the docs describe no field for uploading a font file. If your brand typeface is not on the list, the closest options are a face with a similar feel, plus design.colors and design.typography (weights, scale, stroke width) to bring the look nearer. Latin brand fonts are outside this field altogether, since it covers Hangul faces only.
Sume's docs state that every listed face is SIL OFL 1.1 and ships with the renderer. If your legal team needs more than that, read the licence text yourself; this post is not legal advice.
Sources
Related posts
More in Media tools
- Add a launch date plate to a teaser video with compose overlay
Put a launch-date card on the bottom of a teaser with Timeline compose overlay: width_ratio, margin_ratio, a 300-second ceiling and $0.02 a job.
- Launch teaser music on Lyria 3.5: section markers, 30 seconds
Write a 30-second launch-teaser track on Sume Music Router: section markers, one named moment and no duration field. $0.125 per generation.
- Loop background music under a long video with Timeline 1.0
A 30-second bed under a 90-second video: use soundtrack.loop, fade_out_seconds and duck_db in a Sume Timeline 1.0 render, and plan first for free.
- Mastodon video upload: 99 MB, one video, transcoded to 1300 kbps
Mastodon takes one video per post, up to 99 MB, and transcodes it to H.264 MP4 at up to 1300 kbps. What to send from Sume so the re-encode costs less quality.
Written by Sume