Higgsfield launches Diffuse, an AI image-to-video creator for smartphones

Video artificial intelligence company Higgsfield AI has launched its first smartphone app based on artificial intelligence (AI). The mobile-first app, called Diffuse, is an image-to-video generator that takes selfies and converts them into videos starring characters. The company claims that the characters it creates have lifelike movements. The app is being gradually rolled out to select markets and users will be able to find it on Android and iOS. The AI ​​tool could compete with OpenAI’s Sora, which is reportedly set to launch later this year.

Company official account take Post this announcement to X (formerly Twitter). Higgsfield AI introduces itself as a video artificial intelligence company focused on creating video content for social media. That vision likely comes from the AI ​​company’s CEO Alex Mashrabov, who previously led Snap’s generative AI division. According to the article, the company is building a foundational model for its tool from the ground up that will be entirely video-centric.

in separate postal, the company also launched Diffuse, the first mobile app built by the company, which is currently available in preview mode. “Select a video from the content library, select your selfie, and Diffuse will generate a personalized character in the style of that video. Or use Prompt Builder to create a video from scratch using text, images, or video. Diffuse offers deep personalization , creative control and fine-tuning, so anyone can create exactly what they want (with safeguards, of course),” the post added. The app is currently available in select markets such as India, South Africa, Philippines, Canada, and Central Asian countries.

We were able to find the app in the App Store, but it did not appear in the Android Play Store. This may change in the coming days as the company gradually rolls out the app. The application currently provides 2-second video generation in preview mode. The artificial intelligence company says its ultimate goal is to enable realistic, detailed and smooth video on mobile devices.

To do this, it is building its base model from the ground up. The video model uses the Transformer architecture, which is also used by OpenAI’s ChatGPT. Higgsfield AI further highlighted that it is able to efficiently train its AI models on limited GPUs thanks to a proprietary framework developed in-house. The company has not yet revealed when a full version of the app might be released publicly.