How to document Sora work for client delivery

Sora's shutdown makes archiving a practical priority. Check the export route, retain your original clips and identify which production details you can still recover.

Published July 22, 2026. Updated September 6, 2026.

Sora's web and app experiences ended on 26 April 2026. The Videos API and Sora 2 aliases and snapshots are scheduled for removal on 24 September 2026, following developer notification on 24 March. As checked on 6 September, OpenAI still documents a content export route after the app shutdown. That API date is not a published data-deletion deadline.

OpenAI's Sora 2 system card described C2PA metadata, a visible moving watermark on first-party downloads, internal origin detection and opt-in cameo controls. Those are origin and safety measures, not a complete production account. Inspect the credentials actually present in your archived file, and keep prompts, selected takes, approvals and relevant likeness permissions separately.

Embedded Content Credentials do not disappear merely because the generation service closes. Preserve the original file and its verification result, then connect it to the context your team retained. A file hash helps compare exact copies; by itself it does not establish a creation date, permission or the causal relationship between an input and an output. Mark reconstructed details and unknowns clearly.

What Sora stores on its side

The Sora 2 system card describes C2PA metadata on all assets in its first-party provenance tooling. Source
The Sora 2 system card describes a visible moving watermark on videos downloaded from sora.com or the Sora app. Source
OpenAI described internal detection tools for assessing whether video or audio came from its products. Source
The Sora 2 system card describes explicit opt-in consent and likeness controls for cameos. Source
OpenAI notified developers on 24 March 2026 that the Videos API and the Sora 2 model aliases and snapshots would be removed from the API on 24 September 2026. Source
OpenAI's discontinuation guidance still directs users to sora.chatgpt.com/sunset to request an export. It places permanent deletion after any final export window, without specifying a deletion date. Source

What gets lost without your own record

App closure and data deletion are separate events. Request the available export and inspect its contents before deciding which prompts, iterations or library details need reconstruction.
The visible moving watermark identifies a clip as Sora output, but it says nothing about which account, prompt, or settings produced it.
C2PA metadata lives inside the file, so any downstream re-encode by a platform or an editing pipeline can leave the delivered copy with weaker provenance than your original download.
The system card describes internal detection tools, not a public client verification service. Do not promise a later origin lookup based on that description.
Cameo opt-in controls are not a substitute for checking the permissions relevant to a particular client delivery. Keep the clearance records you actually obtained.

The documentation checklist for Sora

Open sora.chatgpt.com/sunset and select Export; OpenAI says it will email you when the export is ready. Save and inspect the package alongside your delivered clips.
For each client clip, retain available prompts, generation dates, the surface used and approvals. Label estimates or reconstructed details instead of presenting them as original records.
Keep the original download unmodified and store its hash to compare exact copies. Link edited delivery versions separately.
Inspect the Content Credentials of the download with a C2PA verification tool while you still have the original, and file the verification result next to the clip.
If a clip uses a cameo, archive the relevant likeness clearance with the delivery, including whose likeness was used and the permission's scope and date.
Record when you obtained the export and which details it contains, so another team member can distinguish archived evidence from missing context.

Behind The Workflow gives this Sora record a practical home: keep an asset with the prompt, model, references, versions and decisions your team adds. Upload or explicitly capture the available details, fill the gaps, and review the record before handoff. BTW does not automatically reconstruct an external generation history or verify reference rights.

For teams delivering into the EU, the marking inside a Sora file helps with the provider-side machine-readable marking described by Article 50(2). Where a finished image, audio or video qualifies as a deep fake, Article 50(4) separately requires the deployer to disclose that it was artificially generated or manipulated. The Act does not prescribe an internal production log, but a dated project record helps the team show what it generated, what it disclosed and when, even after the originating tool disappears.

Describe what each piece of evidence supports. A verified credential reports its signed assertions. A saved approval records the decision it contains. A hash comparison checks file identity. These are useful together, but none automatically proves every production claim. Keep the original evidence and state which context remains unavailable.

Frequently asked questions

Does the C2PA metadata in a Sora download prove which prompt I used?

Do not infer the prompt from an origin label. Inspect the assertions present in that file, then retain any available prompt history and approval records alongside it. A Sora origin credential is not a complete generation or review log.

Can the provenance signals disappear from my delivered clip?

Re-encoding or other processing can remove embedded metadata. Service closure alone does not remove a credential from an unchanged file. Keep the original download and inspect the credential on the actual delivery copy as well.

Can OpenAI confirm later that a clip came from Sora?

The system card describes internal detection tools without offering a public origin-lookup service. Do not promise clients that OpenAI will verify a clip later. Preserve the original file, its available credentials and your production records.

Sora is being shut down. What happens to the work I already delivered?

Your downloaded clips remain usable. OpenAI still documents content export after the app shutdown; request it now and inspect what it contains. Permanent deletion follows any final export window, and the current guidance gives no exact deletion date.

I used a cameo in a client spot. Is the consent stored anywhere I can show?

The system card describes cameo opt-in controls, but that description does not establish the scope of permission for your client spot. Retain the relevant clearance and check what the actual export contains.

Do Sora's watermark and metadata satisfy the EU AI Act for my delivery?

They help with the provider-side machine-readable marking described by Article 50(2). If the finished content qualifies as a deep fake, the deployer still has a separate disclosure duty under Article 50(4). A delivery record is not prescribed by the Act, but it gives the team evidence of what was generated and how the disclosure was handled.

Test one Sora handoff

Use the original clip, the prompt you retained, and any consent record. The test is whether a teammate can continue from the selected version without reconstructing the session.

Add the chosen output to a BTW project and review its production details. Use upload, the browser extension or a supported connected account, then complete the context this shot needs.
Attach the relevant references and a revised version, then record what changed and which version was accepted.
Review the record as the person receiving the work. Keep missing information explicit and share only the fields agreed for delivery.

Try with one project Download the shot log CSV

Sources

How guides are verified

OpenAI Deployment Safety Hub: Sora 2 Historical description of first-party provenance measures and cameo controls; inspect the actual archived asset.

OpenAI: Sora 2 System Card (PDF) Documents the surfaces Sora was available on and the safety measures around generation.

OpenAI API: Deprecations States that developers were notified on 24 March 2026 of the removal of the Videos API and the Sora 2 aliases and snapshots on 24 September 2026.

OpenAI Help Center: What to know about the Sora discontinuation Current export route and conditional deletion sequence, checked 6 September 2026. The API retirement date is separate from data deletion.

C2PA: Content Credentials technical specification Distinguishes embedded manifests and optional external data retrieval in validation.

Related guides

Start an AI production log Connect one asset to its generation context, revisions and decisions.

Record approval of the right version Keep the reviewer’s decision attached to the exact work they saw.

Prepare a client delivery record Check the agreed handoff, supplied evidence and remaining gaps.

EU AI Act Article 50 for creative teams What the transparency obligations for AI-generated and AI-manipulated content actually say, the real timeline and penalty ceiling, and what a team should record to be ready.

AI provenance: what you can prove and what you cannot An honest map of provenance claims: what a record can demonstrate about inclusion, integrity, and time, and which causal claims no system can support.

Prompt and model documentation: a practical checklist for studios The fields worth capturing for every generation: prompts, models, versions, settings, references, and edits, and how to keep the habit alive in production.

AI film festival submission requirements in 2026 Compare the published WAIFF, AAIFF and Tribeca 2026 rules, check the edition and deadline, and prepare a production file alongside the actual entry requirements.

C2PA vs. AI workflow documentation C2PA and AI workflow documentation solve different parts of the provenance problem. C2PA protects signed assertions attached to media; workflow records preserve prompts, sources, decisions, rights, versions, and approvals around the file.

AI shot log template for image and video production A practical AI shot log template for prompts, models, references, versions, human edits, disclosure, and approval. Download the CSV and adapt the field guide to your production workflow.

The best tools to track prompts, models, versions and approvals for AI work Behind The Workflow is the best tool for keeping prompts, models, versions and client approvals in one record per asset. Air, Frame.io, Ziflow and PromptLayer each lead in an adjacent job; this comparison shows what each records and where each stops.

Guide editorial and verification policy How Behind The Workflow researches, sources, dates, updates, and corrects its AI production guides. Product facts use primary sources, legal claims link to official text, and machine-readable markup mirrors visible content.

How to document Google Veo work for client delivery Document a Veo delivery with its Flow prompt, input frames or ingredients, chosen model, final clip and approval. Learn what SynthID checks add and which production questions they leave open.

How to document Midjourney work for client delivery Midjourney keeps a searchable archive of everything you generate, but the record stays inside your account and images are publicly discoverable by default. This guide covers what midjourney.com actually stores, what never travels with a delivered file, and the per-job checklist that closes the gap.

How to document self-hosted Stable Diffusion work for client delivery A self-hosted Stable Diffusion setup writes its generation parameters into every image by default, at least in the AUTOMATIC1111 web UI most teams run. This guide covers what that buys you, what silently breaks it, and what a client record still needs.

How to document Adobe Firefly work for client delivery Firefly Content Credentials support origin checks. A delivery record adds the prompt, references, model settings, selected version and approval that your team needs to keep separately.

How to document Luma Dream Machine work for client delivery Dream Machine keeps your work organized in Boards and an Ideas stream, but the watermark and the commercial license of a clip depend on the plan that generated it, and the exported file names neither. This guide covers the records that close that gap before client delivery.

How to document Ideogram work for client delivery Ideogram shows the seed and even the Magic Prompt rewrite for every image in its Details Panel, but none of it travels with the downloaded file, and images are public by default. This guide covers what to copy out of the panel and when to generate privately.

How to document Runway work for client delivery Runway keeps your generations in sessions and an asset browser, with a Fixed seed option for related takes, but the video you deliver carries no prompt, seed, or settings. This guide covers what Runway stores in the account and what a producer has to record per shot.

How to document ComfyUI work for client delivery A practical ComfyUI handoff: preserve the graph and dependencies, record the selected render and its reason, then check the files with the receiving artist.

How to document FLUX work for client delivery Save FLUX API results within the ten-minute download window and retain the request, model, seed and approval. Distinguish API logging from Playground workflows.