Together AI today added an application programming interface (API) for generating videos to its platform for building artificial intelligence (AI) applications.
Via the platform, it’s now possible for application developers to mix and match video alongside code, images and voice as needed using an integrated Together Model Platform, says Charles Zedlewski, chief product officer for Together AI.
Developed in partnership with Runware, a provider of an API for invoking image and video models, there are now more than 20 AI models for generating video via the Together Model Platform, including OpenAI Sora 2, Google Veo 3, and Minimax Hailuo.
Additionally, Together AI is now adding support for additional AI models for creating images, including Imagen and Nano Banana from Google and SeeDream from ByteDance.
Based on OpenAI-compatible APIs, Together AI has been making a case for a unified approach to build AI applications to replace what today is a set of fragmented services that each have a unique set of APIs, says Zedlewski.
As a result, developers can also more easily take advantage of the latest AI models to blend text, video, code, image and voice in as little as a minute, he adds. In fact, most developers will soon be using these various AI models with the same application to drive much richer user experiences, he adds. Voice interfaces, especially, might one day soon become a de facto standard for engaging with a much larger percentage of applications than it is today, notes Zedlewski.
There is, of course, now no shortage of options when it comes to building AI applications. The challenge and the opportunity is finding a way to make various AI models accessible to the average application developer without locking organizations into a specific platform.
Hopefully, business applications will become richer as developers master the various APIs being made available. There is no end of examples of how these models are being used on, for example, social media platforms, but the number of these applications that have been deployed with business workflows still remains somewhat limited. However, that is likely to soon rapidly change as developers in enterprise IT environments continue to push the boundaries of AI technologies that with each iteration are rapidly becoming more robust.
In the meantime, however, developers should be experimenting with various AI models, all of which provide varying capabilities at different price points. Just as importantly, developers should be prepared to switch out AI models as the pace of innovation continues to accelerate. “The models will continue to leap frog one another,” says Zedlewski.
There is, after all, no need to repeatedly use the same AI model when a better one is now only a simple API call away.
