Dim busy b-roll so captions read: video filter, then captions, $0.22
Darken a busy clip with video filter dim 0.45 for $0.02, then burn captions for $0.20. Ten clips cost $2.20. The dim check endpoint is free.

Bright, busy footage makes captions hard to read, and the cheapest fix on Sume is to dim the clip before you burn them: video-filter with dim at 0.45 costs $0.02 and a standalone caption job costs $0.20, so $0.22 per clip and $2.20 for ten.
The order matters
Dim first, caption second. The captions run on the filtered MP4, so the white text sits on a darker background and is not itself dimmed. amount is in (0, 1]: 1 changes nothing, and 0.45 reduces the brightness a lot. POST /v1/video-filter/check validates the program for free and returns next_action.
| Step | Per clip | Ten clips |
|---|---|---|
Check (/v1/video-filter/check) | $0.00 | $0.00 |
| Video filter dim 0.45 | $0.02 | $0.20 |
| Standalone captions | $0.20 | $2.00 |
| Total | $0.22 | $2.20 |
Requests
Submit the filter job, wait for the result, then send the result's video_url to captions.
curl -X POST https://api.sume.com/v1/video-filter \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: dim-clip-07" \
-d '{
"video_url": "https://media.sume.com/artifacts/artf_demo/broll.mp4",
"ops": [{ "op": "dim", "amount": 0.45 }]
}'
# then, with the dimmed clip:
curl -X POST https://api.sume.com/v1/video-captions \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: caption-clip-07" \
-d '{ "video_url": "https://media.sume.com/artifacts/artf_demo/dimmed.mp4", "style": "slam" }'Limits and gotchas
- The filter source is limited to 300 seconds.
- Captions need audible speech. For a silent clip send
cueswithtext,startandendinstead. - Korean text on
slam,punchortiktok-greenis rejected (caption_hangul_text_latin_style); use a Hangul style. dimmultiplies luma only: black stays black.
Sources
Related posts
More in Media tools
- Does the Kling motion control prompt change the movement? No
On Sume's Kling 3.0 Motion Control route the motion video drives all movement and the prompt only steers appearance. Fields, limits and $0.1575 per second.
- Eight inspect stills on a 45 s Snap ad: one lands in the first 6 s
Video inspect's default eight stills on 45 seconds sit 5.6 s apart; only 2.8 s falls in Snap's non-skippable first 6. Use frames at 1, 3 and 5.9 instead.
- VEED Fabric on a 41.5-second voice note: $4.20 at 480p, $7.88 at 720p
A 41.5-second voice note bills 42 seconds on Sume's VEED Fabric 1.0: 42 x $0.10 = $4.20 at 480p, 42 x $0.1875 = $7.875 at 720p, reserved as $7.88.
- Fabric accepts image_url or avatar_handle, never both: which to send
On Sume's VEED Fabric 1.0 route, send one visual source: the image_url of an inspected still, or an avatar_handle the user named. Both together is not allowed.
Written by Sume