In the AI‑driven video generation space, OpenAI's Sora 2 app has, since its launch on September 30, 2025, rapidly climbed to the top of the Apple App Store's free apps chart, with cumulative downloads exceeding one million. As an iOS‑only app built on the Sora 2 model, Sora lets users generate surreal short videos with sound from text prompts and images, and includes a TikTok‑style algorithmic feed that supports sharing and interaction. Recently, Bill Peebles, head of Sora at OpenAI, announced on X that the app will soon introduce three major new features to further strengthen creation tools, social interaction, and editing capabilities. This update not only aims to improve the user experience but also marks a shift of AI video from single‑generation toward a social ecosystem. TechCrunch reported that this iteration will help Sora address 'over‑moderation' and pave the way for an Android version, which is expected to be 'coming soon.'
The first of the three new features is 'Character Cameos', a tool that allows users to create cameo characters from real‑world objects or previously generated videos. For example, a user can upload photos or video of a pet dog, guinea pig, or plush toy to generate a reusable AI character and then insert those 'cameos' into future videos. Peebles emphasized: 'You can almost create a cameo for anything, even extracting a generated character directly from a Sora video.' The app interface will also display trending cameos in real time to help users spot popular elements quickly. This feature will arrive within days and marks a natural evolution from static generation to personalized 'upload‑your‑own' experiences. MacRumors noted that the design echoes early communication evolutions—from text to emoji to voice notes—now extending to AI video cameos. To ensure safety, OpenAI requires a one‑time identity verification and grants cameo owners shared ownership: they can delete videos or revoke others' usage rights at any time to prevent unauthorized deepfake misuse.
The second feature focuses on video‑editing capabilities: the initial update will support stitching multiple clips, enabling users to easily combine several short clips into a longer narrative. Peebles said, 'This is just the beginning; more powerful editing tools will follow.' Currently, Sora‑generated videos are limited to 10 seconds, but the editing tools will increase practical utility for marketing shorts and social content creation. The Verge analyzed that this iteration moves Sora from a 'demo tool' toward a 'reliable workflow', likening it to a video industry's 'GPT‑3.5 moment.' Similar to Google's Veo 3, Sora 2 has optimized physical realism and can handle complex motions like Olympic gymnastics or paddleboard backflips, while the editing tools will further bridge AI generation and human post‑production. OpenAI's official blog highlights breakthroughs in spatiotemporal consistency and multi‑shot coherence to ensure stitched videos show no obvious discontinuities.
The third highlight is 'Channels/Groups', designed to build richer social experiences. Users can create private or interest‑based channels—such as sports clubs, university groups, or corporate teams—to share exclusive AI videos. This is similar to TikTok group features but emphasizes familiar networks: algorithmic recommendations will prioritize users' interaction history, location (based on IP), and ChatGPT conversation history (which can be turned off). CNET reported that Sora's feed is 'friend‑oriented' and includes parental controls like limits on infinite scrolling and personalization safeguards to prevent teen addiction. In addition, the app will optimize performance to reduce moderation delays and embed watermarks and provenance signals to ensure content traceability. Engadget added that a Remix function will integrate with channels, allowing users to create variants from trending videos and drive viral spread.
Sora's rapid rise has not been accidental. Since launch it has outpaced ChatGPT and Google Gemini to become the App Store's No.1, with 56,000 first‑day downloads. Appfigures data show invite‑only downloads reached two million, far exceeding expectations. Competition is fierce: Meta's Vibes Feed and Instagram Reels have integrated AI generation, and Sora's 'familiarity‑first' algorithm aims to differentiate but raises privacy concerns—such as IP tracking and portrait‑rights misuse. OpenAI says it has reached a deepfake agreement with the SAG‑AFTRA union prohibiting the creation of public‑figure videos without consent. Currently, the app is limited to iOS users in the United States and Canada (requires iOS 18.0+); ChatGPT Pro subscribers can access Sora 2 Pro via sora.com.
For creators, these features hold significant potential: pet cameos suit playful shorts, the editing tools support professional cuts, and channels help build community ecosystems. Business Standard predicts Sora could reshape social media, with generated content accounting for over 50%. Challenges remain: fake apps proliferated in the App Store (dozens of 'Sora' clones appeared), and OpenAI is pushing for their removal. Looking ahead, an open API and an Android release will expand Sora's reach and support its ambition as a 'world simulator.'
In short, Sora's three upcoming features are more than a technology upgrade; they represent OpenAI's strategic bet on AI‑driven social experiences. With safety as a baseline, the company aims to spark user imagination and shift video from 'viewing' to 'co‑creating.' Despite invite‑only limits, its viral growth shows market demand. Interested users can download the app and join the waitlist to anticipate the next 'GPT moment' for AI video.