Sora capabilities

How Sora is accessed and the schedule on which it is being wound down. Sora was never billed as a fully independent subscription; access historically ran through ChatGPT plans, the standalone Sora app/web, and the OpenAI developer platform. OpenAI has since discontinued the standalone Sora product and removed video generation from ChatGPT, making the discontinuation timeline the single most important fact for anyone planning around it.

Sora's signature consumer feature let a person insert their own verified likeness and voice into generated scenes via a 'cameo'. After a one-time in-app video-and-audio recording to verify identity and capture appearance, the user could appear in Sora videos, and could grant or revoke other people's permission to use their cameo. The same human-likeness capability is tightly restricted in the developer API.

How videos and account content leave Sora, which matters acutely because the product is being discontinued. Generated clips are standard MP4 files, and OpenAI provides a sunset export path so users can retrieve the content they created before each Sora surface goes away.

Sora is OpenAI's generative video product, built on the Sora 2 and Sora 2 Pro models. It turns a text prompt, or a text prompt plus a reference image used as the first frame, into a short cinematic clip with synchronized dialogue, sound effects, and ambient audio generated together with the picture. Sora 2 emphasizes physical realism and shot controllability over earlier video models.

OpenAI's developer Videos API exposes the Sora 2 and Sora 2 Pro models programmatically so applications can generate, extend, and edit video. It is the remaining way to use Sora after the consumer app shut down, but it is itself deprecated with a fixed shutdown date.