Captions with no style set: Korean gets black-outline, no gold tint
If you send no style, Sume picks one from the caption text, and the spoken word gets no gold tint. Set design.colors.active to add one.

If you send no style to Sume video captions, the caption text decides it: slam for Latin text, black-outline for Korean. The default render also has no accent color. The docs say the two default styles carry a gold tint on the spoken word in their original design, but Sume does not add a tint to a look you did not select. Set design.colors.active if you want one.
What you get
Korean text does not resolve to the Latin default, because that would burn Korean as tofu. It also does not resolve to korean-ad. If you want the ad karaoke look, select korean-ad yourself.
| Text | Style resolved | Spoken-word color |
|---|---|---|
| Latin | slam | Fill color, no gold tint |
| Korean | black-outline | Fill color, no gold tint |
| Any, with style set to slam or black-outline | As selected | Style keeps its own gold tint |
Add a tint on a default render
design.colors.active sets the tint in both cases, and Sume merges it over the style's value. This request burns black-outline with a cyan emphasis, and every other value of the style stays.
{
"video_url": "https://example.com/clean.mp4",
"style": "black-outline",
"design": { "colors": { "active": "#22D3EE" } }
}Things to keep straight
languageis a speech-to-text hint. It never picks the style or the font.fontis only for Hangul styles, and a Hangul face with a Latin style returns400 caption_font_requires_hangul_style.- Korean text on
slam,punchortiktok-greenreturns400 caption_hangul_text_latin_style. Sume does not swap the style for you. - The price is the same for any style, so a wrong look is a retry cost, not a discount.
Sources
Related posts
More in Media tools
- Cheapest product video ad: six stills in a Timeline render for $0.30
Six product photos become a 15 s vertical ad with no video model: a $0.10 Timeline render plus $0.20 caption cues. $0.30 against $3.43 for five Wan clips.
- Check an AI pattern tiles seamlessly with ImageChops.offset
Shift a Sume pattern image by half its size with ImageChops.offset and measure the seam stripe. A pass or fail score before you ship a repeating background.
- Check on-screen text in a Wan 3.0 clip: video frames at exact seconds
Wan 3.0 claims text rendering in 12 languages. Before you ship, pull PNG stills at the seconds the text appears with Sume video frames and read them yourself.
- Coffee roaster brand video music: a 40-second lofi bed
Brief a warm 40-second lofi bed with vinyl texture for a coffee roaster brand video: one generation and a one-minute render, $0.225 on Sume.
Written by Sume