Runway Launches Real-Time Video Agent API for Enterprise AI Characters

Joerg Hiller Mar 09, 2026 17:30

Runway Characters API enables businesses to deploy conversational AI avatars from a single image using GWM-1 world model. BBC and Silverside among early partners.

Runway has released Characters, a real-time video agent API that generates fully conversational AI avatars from a single reference image. The tool, powered by the company's GWM-1 world model, requires zero fine-tuning and is now available to enterprise developers at dev.runwayml.com.

The pitch here is straightforward: take any image—photorealistic human, animated mascot, stylized brand character—and turn it into a video agent capable of natural facial expressions, eye movements, lip-syncing, and gestures during live conversations. BBC and Silverside are already using the platform, with broader rollout happening now.

What Characters Actually Does

The API handles the full stack of avatar deployment. Developers define a character with one image, configure voice, personality, and knowledge base through the API, then integrate directly into their products. No model training required.

Where this gets interesting for businesses: the avatars can execute actions mid-conversation. Runway specifically mentions creating support tickets, processing orders from live inventory data, and pulling from enterprise knowledge bases—all in real-time during video interactions.

The underlying tech builds on Runway's General World Models research, which the company has positioned as its approach to understanding and generating coherent visual environments. GWM-1 handles the heavy lifting of maintaining consistent character behavior across extended conversations without quality degradation.

Target Use Cases

Runway is pushing three primary applications. Customer support gets branded video agents that can replace or augment human representatives while maintaining visual identity standards. Learning and development teams can deploy interactive tutors—whether representations of real instructors or entirely fictional guides—for training programs. Marketing teams can finally animate brand mascots and IP characters for real-time customer engagement rather than pre-recorded content.

The enterprise focus makes sense given Runway's trajectory. Video generation has been the company's bread and butter, but the conversational layer opens recurring revenue streams through API usage rather than one-off content creation.

Consumer Access

For non-developers, Runway is launching preset avatars in its web app so users can test the technology directly. The consumer version provides a demo environment while the real monetization target remains enterprise API customers building Characters into their own platforms.

The timing aligns with broader industry movement toward multimodal AI interfaces. Text chatbots dominated the last two years; video-native interactions represent the next battleground for enterprise AI adoption.

• runway ai
• video agents
• gwm-1
• enterprise ai
• conversational ai