Runway has transitioned from being primarily an AI model company to seeking a role as the infrastructure layer for generative media. On Thursday, the company launched the Runway Media Router via its developer platform, Runway Dev, which provides API access to various third-party media models along with its own. The Media Router selects the optimal generation model for image, video, or audio requests based on whether a developer prioritizes quality, speed, or cost, with Runway claiming it is the first router built specifically for generative media.
Anthony Maggio, Runway’s chief product officer, stated that the router aligns with the company’s goal of providing developers with a seamless method to integrate generative media models. The launch, shared exclusively with TechCrunch, aims to facilitate the evaluation process for developers amid a rapid increase in available generative media models. Through Runway Dev, notable customers including Adobe, Cloudflare, ElevenLabs, Expedia, Shutterstock, and Quora can incorporate media generation functionalities directly into their products.
Maggio pointed out that many developers may struggle to grasp the strengths and weaknesses of various generative media models across different outputs. He noted the significant rise in interest among customers for model routing that considers token pricing and quality, especially in light of increasing enterprise expenses related to AI models. The launch follows Runway’s recent shift from unlimited subscription plans to a token-based pricing model, which has drawn criticism from users.
The Router’s intelligence layer reflects the expertise developed by Runway’s in-house creative team in assessing outputs for different media types. This technology was initially designed for Runway’s conversational AI agent product, launched in May, which assists users in creating multi-shot videos and marketing campaigns from text prompts.
Runway’s competitive strategy acknowledges the fragmented generative media landscape, necessitating adaptability to remain relevant. The company’s last AI video model release, Gen 4.5, occurred in December and outperformed similar offerings from Google. Despite recent advancements, Runway’s text-to-video and image-to-video models are no longer leaders in their categories, facing stiff competition from major players like Google and China’s ByteDance and Alibaba.
Instead of requiring developers to commit to a single model, the Media Router assumes that the best model will evolve over time. Co-founder Anastasis Germanidis emphasized the necessity of building a comprehensive stack that includes a developer platform, creative tools, and an inference layer. He noted a growing demand from businesses for Runway’s involvement across all elements of this stack.
“The orchestration matters a lot because people are building entire campaigns with those models,” Germanidis told TechCrunch. He asserted that creating the necessary intelligence layer is crucial for leveraging data from generative models effectively. Maggio added that Runway’s focus since 2018 has been on research and anticipating future developments in the space, ensuring that the company remains at the forefront of generative media innovation.
