Sora app, OpenAI’s new video-making app, quickly rose to the top of Apple’s U.S. App Store. This is a strong sign that people want AI video tools. The app uses the newest Sora 2 model and lets users make short AI videos, remix videos made by other users, and share them on a TikTok-style algorithmic feed.
The launch marks a strategic shift for OpenAI toward a social network based on collaborative experiences. The Sora app has a vertical video feed and swipe-to-scroll navigation that are very similar to TikTok, but all the content is generated by AI. A menu bar on the right side of the feed lets users like, comment on, or remix a video, which encourages people to be creative together.
“Cameos,” an “upload yourself” feature that lets users put themselves into any Sora-generated scene, is a big part of what makes this social experience fun. To use their likeness, users must upload a video and audio recording of themselves that is valid only once. This recording must be very accurate in terms of their voice and appearance.
This feature also lets people share their Cameos with friends, allowing them to use other people’s likenesses in the videos they make. People at OpenAI said this feature was a lot of fun to use and helped them make new friends, suggesting it could be a new and interesting way to talk to each other.
Sora 2, OpenAI’s new state-of-the-art model for making videos and audio, is what makes this experience possible. The model is a big step forward because it can make scenes and sounds that are very realistic. It shows behavior that is more physically accurate and realistic than previous systems.
As OpenAI pointed out, older video models are overly optimistic and may alter reality to match a text prompt. Sora 2, on the other hand, is better at following the laws of physics. For instance, if a basketball player misses a shot, the ball will bounce off the backboard instead of magically moving to the hoop.
The Sora app had a great first day, but it’s currently only available in the U.S. and Canada. OpenAI has said it wants to quickly expand to other countries. This limited access makes its early success even more impressive. The app had 56,000 downloads on its first day and 164,000 installs in its first two days, which is on par with other major AI app launches.
OpenAI is also directly addressing major safety issues with the launch of a social platform. The business gives users clear control over their likeness, letting them choose who can use their cameo, and at any time revoke access or delete any video that uses it. The Sora app also includes parental controls via ChatGPT, which let parents set limits for their teens.
OpenAI will also consider a user’s Sora activity and engagement when making its algorithmic recommendations. The Sora app is designed to promote creation, not consumption, and the company says it doesn’t optimize for time spent in the feed.
At first, the Sora app will be free and have generous generation limits. OpenAI’s only plan for making money right now is to let users pay to make more videos when demand is high. For now, the app is OpenAI’s bet that a social, AI-native video platform is the best way to show off the magic of Sora 2 and start a new era of creative expression.
