Caption a 75-second clip: the $0.20 estimate is for 60 seconds
Sume's standalone caption job lists $0.20 for videos up to 60 seconds. For a 75-second clip, read the live catalog price or split the cut first. Budget table.

A 75-second clip is outside the price that the Sume docs state for a standalone caption job. The docs say each accepted POST /v1/video-captions job reserves and captures $0.20, and that this price is for videos of a maximum of 60 seconds under the current fixed estimate. For anything longer, the docs point you to GET /v1/catalog and the OpenAPI file for the live price. Sume does not publish a per-second caption rate in the docs, so this post does not invent one.
What the docs state, and what they leave open
The table separates the numbers that the docs give from the case that they do not cover.
| Item | Value in the Sume docs |
|---|---|
| Price per accepted standalone job | $0.20 (reserved and captured) |
| Length that the $0.20 is stated for | maximum 60 seconds, current fixed estimate |
| Price for a 75-second clip | not stated; read GET /v1/catalog |
| Input | public HTTPS video URL; localhost, private-network, non-HTTPS and signed/private URLs are rejected |
| Silent clip | fails with caption_no_speech unless you send cues or segments |
Option 1: send the whole clip and read the price first
Before you submit a batch of 75-second clips, call GET /v1/catalog and read the caption entry. That is the price the docs name as the authority. Submit one clip, then look at the usage on the finished job before you queue the rest. This is the only route that keeps the speech in one piece, so the style and phrasing stay continuous.
Option 2: cut the clip at 60 seconds or less
If you want the stated $0.20 price per job, keep each input at 60 seconds or less. Video trim cuts [start, end) out of one Sume-hosted clip into a new MP4 for $0.02 per job. The source can be up to 1800 s and the cut up to 900 s. Use precision: exact (the default) so the cut is frame accurate, and cut at a pause between sentences, not in the middle of a word.
A 75-second clip cut at 38 s gives two parts of 38 s and 37 s. The arithmetic is two trims plus two caption jobs.
| Step | Jobs | Price each | Subtotal |
|---|---|---|---|
| Video trim, part 1 and part 2 | 2 | $0.02 | $0.04 |
| Standalone caption job | 2 | $0.20 | $0.40 |
| Total for the 75-second clip | 4 | $0.44 |
Two checks before you rely on the split
First, the caption job accepts a public HTTPS URL and rejects signed or private URLs. Check that the URL of each trimmed part is one that the caption endpoint accepts. If it is not, host the part yourself at a public address. Second, a cut can split a phrase, so the captions of part 1 end and the captions of part 2 begin at the cut. If the join matters, put both captioned parts back on Timeline 1.0, where a 75-second output bills ceil(75 / 60) = 2 minutes, which is $0.20.
If the clip has no speech, a split does not help. Send cues with text, start and end so Sume burns your text without speech-to-text.
When to choose which
- Choose the whole clip when continuity matters and you can wait for the catalog answer on one test clip.
- Choose the split when your budget sheet needs the stated $0.20 figure for each job and a cut at a sentence break is acceptable.
- Do not multiply $0.20 by 75/60. The docs do not say that the price scales that way.
Sources
Related posts
More in Media tools
- caption_no_speech: why a caption job fails on a silent clip
A Sume caption job returns caption_no_speech when the clip has no audible speech. The fix is cues, segments or words, not a retry.
- Change one caption color: design.colors.active and what it accepts
A Sume caption design override changes one token for one request. Colors take hex, rgb(), rgba() or transparent; other CSS is rejected at request time.
- Captions from a video's speech: $0.20 a job, no SRT needed
Sume's video-captions endpoint transcribes the speech and burns styled captions for $0.20 a job on clips up to 60 s. Optional script_text corrects the wording.
- Check an AI voiceover by transcribing it: a script diff for 4 cents
Run Sume STT on a finished TTS take and diff it against the script to catch misread numbers and names. 4 cents per 450-character line, TTS plus the check.
Written by Sume