Is Sora private? What OpenAI still holds after the shutdown

The Morse teamPublished November 5, 20267 min read

As of November 2026, Sora is not private from OpenAI, and it isn't running anymore. OpenAI closed the Sora app and website on April 26, 2026, and it still holds your Sora videos, prompts, and character recordings until any final export window passes (OpenAI). After that window, OpenAI says it "will permanently delete any data associated with your use of Sora."

We read OpenAI's Sora discontinuation page, its Sora launch and safety posts, its training and provenance pages, and both its current and May 18, 2026 privacy policies on November 5, 2026. The Sora help articles on characters, getting started, and data controls now return "page doesn't exist," so we read OpenAI's own text in archived copies from March and April 2026 and link those. The top results for this question still describe Sora's settings as if the app were live. OpenAI deserves credit for one thing. Its character rules let you see every video anyone made with your likeness, drafts included.

Sora privacy at a glance

Read on November 5, 2026.

Is Sora still available?

No. OpenAI's help center says "The Sora web and app experiences were discontinued on April 26, 2026" and "The Sora API was discontinued on September 24, 2026" (OpenAI). Sora's address at openai.com/sora now sends you to that page, and OpenAI's launch post carries the line "As of April 26, 2026, the Sora product is no longer available" (OpenAI).

The data didn't close with the app. On November 5, 2026, OpenAI's sunset page reads: "Sora is no longer available. You can export your data at any time" (OpenAI). So more than six months after the shutdown, your Sora library is still on OpenAI's servers.

Who can see my Sora videos?

Nobody new, since the feed closed on April 26, 2026. While Sora ran, each kind of data had its own audience.

Who can see what in Sora, then and now

Posting was your choice. OpenAI's safety post says "Videos will only be shared to the feed when you choose to do so," and blocking an account kept it from seeing your profile or messaging you (OpenAI). Sora's first version on the web worked the other way. Its data controls page said "most videos without media uploads are eligible to be published" to the explore feed by default, where others could "remix, blend, or download" them (OpenAI, archived).

Copies outside Sora are beyond any deletion OpenAI does. Every Sora video carried a visible watermark and C2PA metadata, and OpenAI kept "internal reverse-image and audio search tools that can trace videos back to Sora" (OpenAI). Its provenance page warns that "platforms, editing tools, and file conversions may remove" metadata, and as of November 5, 2026 its row for video says only "Sora is discontinued" (OpenAI).

Does Sora use your face?

Only if you made a character, which OpenAI first called a cameo. You recorded "a short one-time video-and-audio recording in the app to verify your identity and capture your likeness" (OpenAI). The help page asked you to read a rotating verification phrase "exactly as written," because "We verify whether the transcript matches" (OpenAI, archived).

"The character feature works by saving your recorded video and using it as a reference point," the same page says, and it "does not store a scan or derive data about you from the video upload." What OpenAI holds, then, is a video of your face and voice.

Three rules shaped who could use it:

• You saw every use. "There is no private use of your character that can't be seen by you," including drafts other people made.
• Changes only worked going forward. Moving from "Everyone" to "Only me" left earlier videos "available for viewing unless you choose to delete them."
• Deleting your Sora account took the videos with it. "If your Sora account is deleted, all videos containing your character will also be deleted."

Did Sora train on your videos?

It could, unless you opted out. OpenAI's training page still says that for "services for individuals such as ChatGPT, Sora, or Operator, we may use your content to train our models" (OpenAI). The archived character page answers the face question directly: "For individual services such as Sora, we may use your content to improve our models unless you opt out" (OpenAI, archived).

Sora's switch lived in Sora's own settings, and OpenAI says "adjusting your settings in the ChatGPT interface will not affect your Sora settings." With the app gone, that switch is gone too. The opt-out that still works is the "do not train on my content" request in OpenAI's privacy portal, which OpenAI says "will apply to Sora too." Our guide to whether ChatGPT trains on your data covers that portal and the ChatGPT side.

Sora also read your ChatGPT account. Its feed page says it "may consider your ChatGPT history" to rank videos, along with "the general location (such as the city) from which your device accesses Sora, based on information like your IP address" (OpenAI). What ChatGPT itself keeps is covered in is ChatGPT private.

Is my Sora data deleted yet?

Not as of November 5, 2026. OpenAI's discontinuation page says that "after the period of time of any final export window passes (if we are able to offer one), we will permanently delete any data associated with your use of Sora," and that it will email you before a final window begins (OpenAI). The page gives no date.

OpenAI's privacy policy puts one limit on deletion. It says deletion doesn't reach content that "has already been de-identified and disassociated from your account" for model training (OpenAI). The Sora page doesn't say whether that applies to Sora content.

The privacy policy itself has moved on. The May 18, 2026 version listed "Sora characters" among the content OpenAI collects, next to files, images, and audio (OpenAI). The current version drops Sora from that list and from its opening line. Our OpenAI privacy policy history dates every version.

How do I export my Sora videos before they're deleted?

Go to sora.chatgpt.com/sunset, sign in with the OpenAI account you used for Sora, and choose Export. OpenAI says "You'll get an email when it's ready," and it recommends you "export your Sora content as soon as possible" (OpenAI).

Once you have the file, keep it somewhere you control. OpenAI's own ChatGPT history is a separate export, which our guide to downloading your AI chat data walks through.

How Morse keeps what you make on your device

Every problem above starts with your prompts, your videos, and a recording of your face sitting on OpenAI's servers, where OpenAI decided who saw them, whether they trained a model, and when they go. Morse is built so your work doesn't live on its servers in the first place.

Morse doesn't make video. Its Studio makes images, and those images are stored in your browser, encrypted, along with your chat history and the characters you make. There's no training switch to find before a shutdown takes it away, because Morse doesn't train AI on your content, sell your data, or show ads. Deleting removes it from your browser, and Morse keeps no other copy. Clearing your browser or deleting your account removes your images and chats for good, and Morse can't recover them, so keep a copy of anything you need.

For the chats around your work, pick a mode per chat:

• Private, the default, for scripts and ideas you'd rather no host keep. Open-weight models like DeepSeek, Kimi, and GLM, plus xAI's Grok, run with your identity stripped, and Morse has zero data retention agreements with its inference, search, and other providers.
• Anonymous, to use GPT, Claude, or Gemini without the company knowing it's you. Morse's proxy removes your account, email, and IP address. The provider still sees the words and may keep them for a limited time to check for abuse, under its own policy, so leave out names and details that point to a real person.
• E2EE, for anything no one should read. Your browser checks the enclave's proof that it's genuine, then seals the message for that enclave alone, so Morse's servers relay text they can't read. It rests on the enclave hardware and attestation working as designed, and it runs open-weight models without image generation.

Whatever the mode, your history stays in your browser, locked when you sign out, and Morse usually has very little to give if a court asks. If what you liked about Sora was making pictures of ideas, our guide to Midjourney alternatives compares image tools on who can see what you make.

Try a chat from the Morse home page without an account, or start free with 3 Studio images a day and no card needed.