Runway is trying to make itself less dependent on having the flashiest AI video model and more central to how companies build generative media products.
The startup launched Runway Media Router on Thursday through Runway Dev, its developer platform for companies using image, video and audio models through APIs. The tool is designed to choose a model automatically for each request, depending on whether a developer wants better quality, faster output or lower cost.
Runway says model routing has already become familiar in text-based AI, where companies switch between large language models to manage performance and token bills. The company says its new router is built for media generation, where the choice is no longer just about text accuracy, but about motion, composition, audio and lip syncing.
Anthony Maggio, Runway’s chief product officer, said the aim is to make the platform the “easiest one-stop shop” for developers adding generative media to their own products.
The launch follows Runway Dev’s rollout earlier this month. Through that platform, customers including Adobe, Cloudflare, ElevenLabs, Expedia, Shutterstock and Quora can use Runway’s API rather than sending users to Runway’s own website or app.
The timing matters because generative media is no longer a one-company race. Google, ByteDance, Alibaba and other companies are pushing video and image models, while developers face a growing menu of tools with different prices, policies and performance levels.
Runway executives said developers may also want to set preferences around where models come from, including whether to use Chinese model providers, as Washington weighs restrictions on some Chinese open AI systems.
For studios, app makers and marketing teams, the development could reshape a practical question: not which AI model wins the leaderboard today, but which system quietly chooses the right one tomorrow.
