You can't use Sora 2 anymore: OpenAI's own page says the Sora product has been unavailable since April 26, 2026. Sora 2 was OpenAI's video and audio model behind a social iPhone app. Similar results now come from other AI video tools, and the prompting habits still carry over.
If you've landed here looking for steps to make a video in Sora 2, I'll save you some frustration straight away: OpenAI has closed it. Its own Sora 2 page now says the Sora product has been unavailable since April 26, 2026. Plenty of guides online still describe it as if it's open, which is why so many people are still searching.
So this page does three useful things instead. It explains what Sora 2 was and how it worked, what OpenAI said about safety and consent, and how to get similar results with the AI video tools you can still use, with the habits that carry over.
What Sora 2 was
OpenAI described Sora 2 as its flagship video and audio generation model. Unlike earlier AI video tools that made silent clips, it could create speech and sound effects alongside the pictures, so a short scene could come with dialogue and background noise.
Sora 2 powered a social app for iPhone, simply called Sora, where people could create videos, remix other people's clips and scroll through a feed. OpenAI first rolled it out in the U.S. and Canada as an invite-based app, so you came in alongside friends.
How Sora 2 worked when it was open
- You typed a description of the scene you wanted, much like a prompt in ChatGPT.
- Sora generated a short video with matching audio.
- You could remix a video from the feed, changing the prompt to take it somewhere new.
- You could create a character based on your own likeness and choose who else was allowed to use it.
The character feature, often called cameos in news coverage, was the part that got the most attention. It let you appear in AI videos, with controls over who could put you in them.
What OpenAI said about Sora 2 safety and consent
OpenAI published a page on launching Sora responsibly. The points worth remembering, because the same questions apply to every AI video tool, were these:
- Consent: OpenAI said only you decide who can use your character, you can revoke access at any time, and you can always see videos that include your character, even other people's drafts.
- Provenance: it said every Sora video included visible and invisible signals showing it was AI-made, including a visible watermark at launch and C2PA metadata, an industry-standard signature.
- Well-known people: it said it blocked depictions of well-known people unless they used the characters feature themselves.
- Younger users: it said Sora included stronger protections for teens, including limits on mature content.
Can you still use Sora 2?
No. OpenAI's own page is clear that the Sora product is no longer available. If you see a website, app or advert offering 'Sora 2' access, treat it with real caution: it isn't OpenAI's product, and some lookalikes exist to collect sign-ups or card details.
It's a useful reminder for anyone building a business on a single AI tool: products can change or close quickly, so keep your own copies of anything you make and avoid depending on one app for important content.
If you made videos in Sora before it closed and want to know whether you can still get them, check OpenAI's help centre for the latest guidance rather than relying on third-party advice.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
What to use instead of Sora 2 for AI video
The good news is that AI video has plenty of alternatives. Google's Veo, Kling, Runway and several others make short clips from text or images, some with sound. If you're moving to one of those, my Veo 3 prompts and Kling AI prompts give you ready-made starting points.
When choosing a replacement, I'd look at these things rather than hype:
- Whether it creates audio and dialogue, or only silent video.
- How long each clip can be and whether you can extend it.
- Whether it can start from your own image, such as a product photo.
- What its terms say about commercial use of the videos.
- How it labels AI-generated content and protects people's likenesses.
Prompting habits from Sora 2 that still work
- Describe the shot like a film director: subject, setting, camera angle and movement.
- Mention lighting and mood, such as soft morning light or a busy evening street.
- Say what sound you want, if the tool supports audio.
- Keep each clip to one simple action rather than a whole story.
- Generate a few versions, then refine the prompt that came closest.
Using AI video in a small business after Sora
For most small businesses, AI video works best for short social clips, product teasers, simple explainers and background footage, not for anything pretending to be real. My guide to generative AI video tools for business goes through the kinds of jobs they suit.
- Only use photos and footage you own or have permission to use.
- Never make videos of real people without their written consent.
- Label AI-generated video where a platform asks you to.
- Check each tool's rules before using clips in adverts.
Frequently asked questions
Is Sora 2 still available?
No. OpenAI's Sora 2 page says the Sora product has been unavailable since April 26, 2026.
What was Sora 2?
Sora 2 was OpenAI's video and audio generation model. It powered a social iPhone app called Sora, where people created and remixed short AI videos with sound.
How did cameos work in Sora 2?
You could create a character from your own likeness and choose who could use it. OpenAI said you could revoke access at any time and see videos that used your character.
What can I use instead of Sora 2?
Other AI video generators such as Google's Veo, Kling and Runway create short clips from text or images. Check each tool's terms and safety rules before you use it.
Is a website offering Sora 2 access safe?
Be cautious. OpenAI has closed Sora, so any site or app claiming to sell Sora 2 access isn't OpenAI's product.
