Transcript of a 10-minute client video before localizing it: cost
Sume video inspect transcribes at $0.01 per audio minute, with a 600-second hint cap, so 10 minutes is $0.10 plus compute. Set language_code, get sentences.

Ten minutes of audio is $0.10 at the transcript rate on Sume video inspect: $0.01 per audio minute. Inspect also bills its Modal compute for probing, so the call costs more than the transcript line; the docs do not give one all-in number, and they say every inspect reserves its compute ceiling and is captured at what actually ran, never above the hold.
Settings that matter for translation work
- transcribe: true turns it on. Without it, language_code, segmentation and duration_seconds return 400 video_inspect_transcribe_required.
- language_code is a hint such as en or ko. Omit it for auto-detect.
- segmentation.mode sentence also returns gapless sentence segments, which are caption-line shaped, with silence_split_seconds from 0.2 to 3.
- duration_seconds is a reservation hint with a maximum of 600. Omit it and Sume reserves 1 minute.
- The source clip can be up to 1800 seconds.
Cost by length
| Audio length | Hint to send | Transcript rate |
|---|---|---|
| 3 min | 180 | $0.03 |
| 10 min | 600 | $0.10 |
| 20 min | Hint is capped at 600 | Transcribe in 10-minute clips |
| Silent clip | n/a | inspect_source_has_no_audio |
Before you pay
Run a probe-only inspect with frames: false first. Check probe.has_audio, because a transcribe request on a clip with no audio track fails as inspect_source_has_no_audio. A translator then works from words with times and sentence segments rather than from a raw audio file.
Request
curl -X POST https://api.sume.com/v1/video-inspect \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: transcript-1" \
-d '{"video_url":"https://media.sume.com/artifacts/artf_demo/talk.mp4","frames":false,"transcribe":true,"language_code":"en","duration_seconds":600,"segmentation":{"mode":"sentence"}}'Sources
Related posts
More in Media tools
- TikTok ad caption budgets: 100 in TopView, 150 for Spark Pull, 4 lines
TikTok's TopView page caps ad captions at 100 characters, Spark Pull allows 150, and both display four lines. Limits by placement and short burned-in text.
- TikTok ad profile photo: 98x98, under 50 KB, logo in the 66px centre
TikTok in-feed ads use a 98x98 square profile photo under 50 KB, with key elements inside a 66px centre. Pull a brand still with Sume video frames and fit it.
- TikTok ad safe zone shrinks with caption length: place burned-in text
TikTok says the ad safe zone changes with format and caption length. How to place burned-in captions inside it with Sume's caption anchor_ratio and frames.
- TikTok Automotive Ads specs: 4:3 800x600 images, 5 to 60 s video
TikTok Automotive Ads use 4:3 product images (800x600 preferred, 450x450 minimum) and 5 to 60 second video. Make an 800x600 clip with Sume Timeline.
Written by Sume