An OpenRouter-compatible video API: sume/auto or a pinned model

Sume's POST /v1/videos follows OpenRouter's video generation API field for field. Let sume/auto pick the model, or pin a catalog id like seedance-2.5.

Sume's video generation endpoint, POST /v1/videos, follows the OpenRouter video generation API field for field. A client written against OpenRouter's docs works against Sume after you change the base URL and the API key. You can pin a model from Sume's catalog, or send model: "sume/auto" and let Sume pick.

Every detail below is from the Video generation docs.

How does a video job work?

Video generation is asynchronous because it takes much longer than text or image generation.

Submit a request to POST /v1/videos.
Receive a job id and a polling URL immediately (202).
Poll GET /v1/videos/{jobId} until the status is completed, or pass callback_url and receive a signed webhook.
Download the result from GET /v1/videos/{jobId}/content.

curl -sS -X POST "https://api.sume.com/v1/videos" \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: teaser-001" \
  -d '{
    "model": "sume/auto",
    "prompt": "A vertical UGC-style product clip on a desk, natural light",
    "aspect_ratio": "9:16",
    "duration": 5
  }'

Should I use sume/auto or pin a model?

Use sume/auto when you want Sume to choose the model family. Routing is a pure function of the normalized request and the catalog version, so an idempotent replay prices and routes the same way. Responses report sume/auto; Sume does not disclose which family served the job, so do not build on traits of the output to infer it. Per the Video Router docs, Auto create controls default to 720p and 8 seconds, with 3–10 second clips at 16:9 or 9:16.

Pin a catalog id when you need a specific model's behavior or limits. List the catalog with GET /v1/videos/models; each entry carries its supported resolutions, aspect ratios, durations, and pricing SKUs.

Example per-model limits from Video generation, 2026-09-23. Check `GET /v1/videos/models` before you submit.
Model id	Duration	Resolution
`seedance-2.5`	4–30 s	480p, 720p, 1080p
`wan-3.0`	2–30 s	See catalog
`minimax-h3`	5–15 s	Native 480p, 768p
`gemini-omni-flash-1.1`	3–10 s	360p, 720p, 1080p, 4K
Every other catalog model	Up to 15 s	See catalog

Where does Sume differ from OpenRouter?

The docs keep one table of every delta, in the Sume differences section. The ones you notice first:

Base path is https://api.sume.com/v1/videos, with no /api segment.
Model ids are bare catalog ids such as seedance-2, never org/slug.
sume/auto is a Sume-only addition.
Send Idempotency-Key to make retries safe; a replay returns the original job.
Webhooks use Sume's standard job envelope, signed with x-sume-webhook-signature.

How is it billed?

Jobs draw from your workspace USD balance, reserved on submit at the provider's list rate × 1.25, on every model. Plans and rates are on Pricing and API pricing. Pro and higher plans list access to 100+ video generation models.

Is there the same thing for music?

Yes. The Music Router picks the engine when model is omitted or set to sume/music-auto (Lyria 3.5 today), and passes explicit catalog ids such as lyria-3.5 through to that engine.