Video upscale API: a 12-second clip is $0.108, the cap is 30 seconds
Sume Video Upscale 1.0 bills $0.009 per input second, up to 30 seconds ($0.27). Factor 1.1 to 4, three tiers, and a 5-second default reserve.

Sume's Video Upscale 1.0 charges $0.009 per input second and accepts up to 30 seconds. A 12-second clip costs 12 x $0.009 = $0.108, and the longest clip costs $0.27. The route is POST /v1/video-upscale-1.0/upscale, and Sume picks the provider model server-side.
Request fields
The body is strict: unknown keys are rejected, and sending model, endpoint, or provider_endpoint returns an error, because the model is chosen server-side and the public id is sume/video-upscale-1.0. The source must be a public HTTPS video URL.
| Field | Values |
|---|---|
| video_url | Public HTTPS video URL (required) |
| upscale_factor or scale_ratio | 1.1 to 4 |
| duration_seconds | 1 to 30; if omitted, Sume reserves 5 seconds |
| enhancement_tier | fast, standard, or pro |
| metadata | Free-form object |
The arithmetic
The catalog describes the price as per input second after Sume margin: $0.009. If you omit duration_seconds, the reservation is 5 seconds, which is $0.045; then pass the real length so the hold matches the clip. For 12 seconds: 12 x 0.009 = $0.108. For 30 seconds: 30 x 0.009 = $0.27. A 31-second video is over the cap, so split it first. The catalog lists one per-second rate for this route and does not list a separate price for each tier, so confirm the live catalog before you assume a tier is free of extra cost.
curl -X POST https://api.sume.com/v1/video-upscale-1.0/upscale \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"video_url": "https://example.com/clips/12s.mp4",
"duration_seconds": 12,
"upscale_factor": 2,
"enhancement_tier": "standard"
}'Job handling
The route returns a job. Poll GET /v1/jobs/:id/status and GET /v1/jobs/:id/result, or register a webhook. The usual communication fields (mode, webhook_url, wait_timeout_seconds) apply. Source: the Sume catalog (read 2026-10-08) for the rate and the repo API schema for the field limits.
Cost table
Because the rate is linear in input seconds, you can price a batch in your head. Use the table as a reference, with the catalog rate of $0.009 per second (read 2026-10-08). The 5-second row shows what an omitted duration_seconds reserves.
| Input seconds | Arithmetic | Cost |
|---|---|---|
| 5 (default reserve) | 5 x 0.009 | $0.045 |
| 12 | 12 x 0.009 | $0.108 |
| 20 | 20 x 0.009 | $0.18 |
| 30 (cap) | 30 x 0.009 | $0.27 |
Webhooks and polling
Upscaling is a provider job, so it is queued and takes longer than a worker ffmpeg step. Register a webhook_url, or poll the job status endpoint, rather than holding a connection. If you submit a batch of ten 12-second clips, the cost is 10 x $0.108 = $1.08, and the jobs run independently, so one failure does not affect the others. See the jobs guide for the status model.
Sizing the job
Pass the real clip length in duration_seconds so the hold matches the clip, and keep the factor modest: the allowed range starts at 1.1 and stops at 4. For a clip longer than 30 seconds, trim it into pieces under the cap first with video trim and upscale each, for example a 60-second clip as two 30-second jobs at $0.27 each, $0.54.
Sources
Related posts
More in Media tools
- Vidu Q4 540p draft to 1080p: Sume video upscale at 2x costs $0.09
A 10-second 960x540 draft upscaled at the default factor 2 gives 1920x1080 for $0.09 on Sume. Sume does not list Vidu; the clip only needs a public HTTPS URL.
- Voice greeting on a still photo: TTS first, then H3 Max lip sync
To make a still photo speak, make 5 to 14.8 seconds of TTS audio, then call MiniMax H3 Max lip sync. TTS costs $0.0475 per 1,000 characters.
- Voiceover plus music bed: TTS and Lyria cost, duck_db in Timeline
A 700-character voiceover ($0.03325) and one Lyria bed ($0.125) cost $0.15825 before the render. Timeline duck_db (0 to 20) lowers the bed under speech.
- Which video specs TikTok, LinkedIn, Pinterest and Shorts leave blank
Only TikTok lists a bitrate; only LinkedIn lists a frame rate; Pinterest lists no resolution; the Shorts page lists no size at all. What Sume can set for each.
Written by Sume