Korean BBQ restaurant reel: korean-ad captions on spoken Hangul

To caption a restaurant reel with spoken Korean, send the clip's public HTTPS URL to POST /v1/video-captions with style: "korean-ad" and language: "ko". Sume describes korean-ad as a CapCut-style Hangul karaoke look for Korean speech: one short phrase at a time, in the lower third, with the spoken word changing to a heavy weight.

Do not use slam, punch or tiktok-green for Korean text. Those styles use Latin display faces without Hangul glyphs, and Sume rejects the request with 400 caption_hangul_text_latin_style rather than burn tofu boxes onto your video.

What happens if you send no style

If you omit style, the text of the captions sets it. Korean text resolves to black-outline, which is a safe default: white fill on a thick black outline, in the middle of the frame. It does not resolve to korean-ad, so if you want the ad look with the accent color on the spoken word, say so.

language only tells speech-to-text which language to expect. It never selects the style or the font.

The request

The clip must have audible speech. A silent clip fails with caption_no_speech, and the next action is to use overlay captions. For a menu reel with music only, send authored cues with text, start and end instead, and use a Hangul style for them.

If you have the exact script, pass it as script_text to align the captions to your words rather than to the transcript, which helps with dish names.

curl -X POST https://api.sume.com/v1/video-captions \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: bbq-reel-captions-001" \
  -d '{
    "video_url": "https://example.com/bbq-reel-clean.mp4",
    "style": "korean-ad",
    "language": "ko",
    "script_text": "오늘 저녁은 숯불 갈비로 정했어요"
  }'

Pick a font and a look

font is optional and only applies to Hangul styles. Pretendard is the face for korean-ad, and Do Hyeon for black-outline and clip-wipe. Setting font changes the face and keeps the style. design overrides colors, typography, placement, phrasing and motion for one request, for example a brand-red active word.

Where the speech comes from

If your reel has no voice, Sume's avatar talking-video route can produce a spoken Korean-script clip with inline captions, and a Korean script on a Latin style is rejected there too with the same error. Check the final frames on a phone at arm's length: menu names should be readable without zooming.