OpenAI is fine-tuning a TikTok-style app focused exclusively on clips created by artificial intelligence. According to several reports, the platform integrates a vertical feed and a recommendation system for users to discover videos generated with Sora.
The proposal, still in the testing phase, aims for simple use: short videos of up to 10 seconds, social tools, and options to remix other users' creations. All content must be AI-based, so You will not be able to record with the camera or upload files from the gallery.
What this new platform would offer
According to the available information, the app replicates the most familiar mechanics of short formats: a infinite scroll for recommendations and quick actions on each clip. Curation is based on an algorithm that learns from the user interactions.
- Videos generated with Gravel 2 and maximum duration of approximately 10 seconds.
- Social interactions: I like it, comments and remix option.
- No external loads: does not support upload photos or videos nor record with the mobile phone camera.
- Function of identity verification; if passed, it is possible to use your own image in the creations.
- Development as standalone app focused on Sora 2.
How to create and share clips
The generation part of templates and templates to transform ideas into short videos. Additionally, you can remix already published content to explore variations and styles, an approach that facilitates quick creativity without leaving the feed.
The experience is designed so that everything passes through the AI “sieve”: there is no direct recording option nor to attach materials from the reel. With this, the platform maintains a clear editorial line: a catalog composed only of synthetic content.
Project status and schedule
Sources close to OpenAI indicate that is already testing it internally with employees. First impressions have been very favorable and using the feed has become part of their routine, although there is no release date confirmed nor details of availability by region.
For now, no final designs or information about advanced features have been shared. What does seem clear is the direction: fast consumption of clips, social participation and native generation with Sora as the creative engine.
Technical limitations and safeguards
The current version of Sora has challenges that OpenAI will have to polish for a more robust experience: it has trouble respecting certain physical laws, maintain coherent spatial relationships (e.g., distinguish left and right) and follow a consistent narrative in longer sequences.
In parallel, the company applies controls to prevent improper use, with copyright filters and restrictions on protected or potentially sensitive content. The goal is to minimize legal risks and strengthen the ecosystem security since the first day.
Competition and market context
The movement occurs in the middle of the race for the AI-generated videos. Meta has introduced Vibes as its own showcase for synthetic clips, while the popularity of ChatGPT could drive adoption of OpenAI's venture. Added to this is the uncertainty surrounding TikTok in the United States, a scenario that opens the way for new actors.
If the proposal is confirmed, the OpenAI app would be able to compete head-to-head in the most popular format of the moment: vertical, short and highly viral videoIt remains to be seen what its roadmap is, how Sora's capabilities evolve, and what tools it offers for creators and brands.
With what is known so far, everything points to a platform that combines algorithmic discovery, social interaction and native generation with Sora 2, avoiding external uploads and prioritizing the coherence of the catalogue; a sharp approach that, if combined with technical improvements and robust safeguards, could make it a significant new player in the short-form video world.