C2PA inputTo records prompts and seeds; Sume video has no seed
The C2PA guide says prompts, reference images and seed values can be recorded as inputTo ingredients. Sume's v1 video models accept no seed, so log other facts.

The C2PA implementation guide names inputTo ingredient assertions as the place to record inputs to a generation: prompts, reference images, seed values and parameters. On Sume you can record the prompt, the reference URLs and the model id, but not a seed, because the Sume docs say no v1 video model accepts seed.
What C2PA asks for
The guidance says prompt and model information can be recorded as ingredient assertions. The implementation guide extends that to reference images, seeds and parameters. These records make a generation more auditable, since a reviewer sees what went in.
| Input | Possible on Sume video? |
|---|---|
| Prompt text | Yes, you keep what you sent |
| Reference image or video | Yes, via input_references where the model supports them |
| Seed value | No, no v1 video model accepts seed |
| Model id | Yes, unless you used sume/auto |
Replace the seed with a replay record
Without a seed you cannot promise the same pixels twice. What you can keep is the request body and the result. Submits require an Idempotency-Key, which means a retry of the same key maps to the same job, not a second generation.
Store the key with the request body. Re-sending the same key is a retry of that job, not a re-roll.
Billing note
Every video model bills at provider list times 1.25, reserved on submit. A re-roll is a new paid job, so keeping the approved job id is cheaper than regenerating to prove a point. See video models.
Sources
Related posts
More in Developers
- C2PA Interim Trust List frozen: why Content Credentials look untrusted
The C2PA Interim Trust List was frozen on 1 January 2026. What that means when a validator flags credentials, and what to check on files you deliver.
- C2PA regions of interest: mark only the AI segment of a video
The C2PA implementation guide covers regions of interest for audio and video segments, and the 2.3 spec allows time ranges. What that enables for mixed footage.
- C2PA soft binding and watermarks: what a trim or re-encode does
The C2PA 2.3 spec defines soft binding for invisible watermarks. Sume's trim has exact re-encode and keyframe copy modes; here is what to verify after each.
- Can polling Sume job status hit the rate limit? Read budget math
Reads and writes have separate Sume budgets. Worked numbers for polling every accepted job per plan, plus a delay function that reads ratelimit headers.
Written by Sume