Choosing an AI video API as a developer: a technical checklist
Choose an AI video API on job model, not demos: async submit and poll, idempotent retries, webhooks, live capability discovery, a billing model you can read

When you pick an AI video API for a product, check the plumbing before the clips: is it asynchronous, can you retry safely, can it call you back, can you discover model limits at runtime, and is the price reserved before the work starts. This is a checklist, not a ranking. Here is how Sume's /v1/videos answers each item, per its docs read 2026-09-29.
Source: Video generation.
What should the checklist cover?
| Question | Sume `/v1/videos` |
|---|---|
| Async job model | POST /v1/videos, then GET /v1/videos/{id} and /content |
| Safe retries | Idempotency-Key header; a replay returns the original job |
| Webhooks | callback_url (HTTPS), signed with x-sume-webhook-signature |
| Capability discovery | GET /v1/videos/models |
| Wire shape | OpenRouter video shape, bare model ids |
| Cost visibility | Reserved on submit at provider list × 1.25; usage.cost on the job |
| Unsupported fields | size, seed, provider.options return 400 unsupported_parameter |
How do I submit and poll?
curl -X POST https://api.sume.com/v1/videos \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: job-123" \
-d '{"model":"sume/auto","prompt":"A sailboat at dawn","duration":5}'
curl https://api.sume.com/v1/videos/$JOB_ID \
-H "Authorization: Bearer $SUME_API_KEY"
What else should I test?
- Failure paths: send an unsupported value and read the error.
- Retry with the same
Idempotency-Keyand confirm you get one job. - Poll until a terminal state and time-box the wait.
- Check the price of your real clip before scaling.
Do I need to pin a model?
Not to start. sume/auto lets Sume pick; pin an id once you need a specific input. See sume/auto or a pinned video model.
Sources
Related posts
More in Models
- AI video with audio: which models always add sound, which are optional
On Sume, minimax-h3, minimax-h3-max and gemini-omni-flash-1.1 always add native audio; kling-3, Seedance and Wan make it optional; grok-imagine-video-1.5 none.
- sume/auto or a pinned video model: how to decide
Use sume/auto when you accept 3 to 10 second clips at 16:9 or 9:16 and do not care which family runs. Pin a model id for a specific input, length or audio.
- Voice cloning API: clone once in the app, speak via the API
Sume has no API route that creates a voice clone. You clone once in the app, then send the voi_ id to the text to speech API on every call.
- Wan 3.0 1080p: how to request full HD from the API
Wan 3.0 outputs 480p, 720p or 1080p. Send resolution 1080p to wan-3.0 on POST /v1/videos, up to 30 seconds, with audio. See the request and the price.
Written by Sume