Caption line breaks: Netflix bottom-heavy pyramid rule vs Sume cards
Netflix wants two-line captions broken as a bottom-heavy pyramid. Sume caption cards break by phrasing settings. How to approximate the rule and where it ends.

The rule
The Netflix English Timed Text Style Guide recommends line breaks that form a bottom-heavy pyramid when a caption runs to two lines, meaning the second line is the longer one. The guide pairs this with 42 characters per line and a limit of two lines.
The reason is readability: a longer top line pushes the eye across a wide span and then down, while a shorter top line feels balanced.
How Sume cards break
Sume's burned-in caption styles group words into phrase cards using design.phrasing: max_words, max_chars and pause_seconds. Cards are split by those settings and by pauses in speech, not by a pyramid rule. You can supply authored cues, where each cue's text may include a newline for a two-line card, which gives you exact control over where the break goes.
| Input path | Who decides the break |
|---|---|
Speech with phrasing settings | Sume, by words, characters and pauses |
Authored cues with a newline | You |
editorial-emphasis style | The style drops the phrase-final word to a second line |
Writing pyramid breaks yourself
With authored cues you can apply the rule in code. Take the text, find the split nearest the middle that keeps the second line equal or longer, and prefer to break after punctuation or before a conjunction.
- If the whole phrase fits on one line of 42 characters, do not break it.
- Otherwise split near the middle and move the break so line two is at least as long as line one.
- Do not split names, numbers with units, or verbs from their particles where you can avoid it.
- Send the result as
cueswith a newline intext.
What to leave alone
The editorial-emphasis style is a deliberate design: its second line is a larger phrase-final word. That is a stylistic choice for ads and is not the Netflix convention. Do not compare it to the guide.
The aim here is a style rule, not a certification. If you deliver to a broadcaster, follow their spec and check with their tools.
A small helper to place the break
For a phrase of up to 84 characters, try every space as a break point and score it: line two must be at least as long as line one, neither line may exceed 42, and the difference in length should be small. Pick the lowest-scoring break and add a bonus when it follows a comma.
Run it over your cue list, send the result as cues, and keep the originals so you can compare the look against Sume's own grouping.
Takeaway
Sume decides breaks from speech and phrasing settings. If you want a bottom-heavy pyramid, write the cues yourself and pass a newline where each break goes.
Sources
Related posts
More in Media tools
- Captions for applause and music cues: author the cue text yourself
ElevenLabs Scribe can tag laughter and applause. On Sume, a silent clip returns caption_no_speech, so supply your own cues.
- Check a product-swap video edit for the old product with frames
After a prompted product swap on a video, pull matching stills from the source and the edit and compare them to catch the old product. Sume docs for each step.
- Check a video crop before you pay: free video-filter /check
Validate a crop or dim program with POST /v1/video-filter/check, which is free, then submit the $0.02 encode. Diagnostics, refusal codes and next_action.
- Check the transcript before captions: Scribe v2 error tiers
ElevenLabs lists Scribe v2 error rates by language group. Before burning captions on Sume, pass script_text or your own words so a wrong transcript never ships.
Written by Sume