Runway launches Runway Agent, its conversational mode.

Runway launches Runway Agent, a conversational tool extending Gen-4 and Gen-4.5 models to manage end-to-end video production from a natural language brief.

Runway introduces Runway Agent, a conversational mode that handles the production of a complete video, from a natural language brief to an edited, sound-designed, and ready-to-publish deliverable. The tool extends the Gen-4 and Gen-4.5 models, moving from generating isolated clips to managing an end-to-end video workflow.

The typical workflow: the user describes their intent (for example, '15-second Instagram ad for premium headphones in a mid-century modern style'), can attach reference images to frame a character, product, or setting, then selects an aspect ratio, duration, and audio preferences. The agent returns a concept along with a narrative structure broken down into shots, with descriptions of camera movements, lighting, and framing. The exchange continues in dialogue to adjust the tone, modify the text, remove, or add a scene. Once the framework is validated, the agent generates the multi-shot sequence in high resolution with integrated voice-over, music, and sound design, then places everything into a timeline editor for final adjustments.

The use cases highlighted at launch cover advertising, short-form social content (Instagram, TikTok, YouTube), product videos, previsualization for film and television, as well as animated or stylized content. Runway Agent is immediately accessible via browser at app.runwayml.com/agent. A launch offer provides 50% off the Pro subscription with code RUNWAY50, valid until Monday, May 18, 2026.