Runway lets creators animate multiple video elements at once

Making AI video has long felt like spinning a roulette wheel. You upload a still image, hit generate, and hope the whole scene doesn’t warp into a blurry mess.

Runway wants to change that. The New York-based startup just rolled out Multi Motion Brush to all users of its Gen-2 foundation model, bringing independent motion controls to different parts of the same image.

The feature was previously locked behind Runway’s Creative Partners Program for power users. Now, it is officially live for everyone on Gen-2, marking a first among commercially available AI video generators, where rivals still limit motion to the entire frame or just one isolated spot.

Directing characters and backgrounds separately

Runway originally debuted a single Motion Brush in November 2023. It worked, but you were stuck applying one universal movement pattern per shot.

Multi Motion Brush fixes that bottleneck. The workflow starts with uploading a static image prompt. From there, you use a cursor brush to paint distinct objects in different digital colors.

Each colored mask gets its own independent set of slider controls in the web interface. You can set clouds in the sky drifting left while sending a walking subject to the right, adjusting horizontal, vertical, and proximity axes.

The sliders move from -10 to +10 with decimal precision, allowing fine-tuned pacing for left/right, up/down, or zoom in/out dynamics. If a pass goes sideways, a single Clear button resets everything back to zero.

The update adds to the 30-plus creative tools Runway has layered onto Gen-2 since its March 2023 debut. The platform previously expanded clip lengths from 4 to 18 seconds in August and added a Director Mode for camera pans and tilts.

Still, the underlying engine has limits. Runway’s outputs can still run into motion blur, incomplete limbs, and temporal flickering.

Why it matters

For digital creators and video editors, this tool is worth trying right away. Instead of spending hours masking, keyframing, and compositing separate AI layers in After Effects, you can now orchestrate multiple elements directly inside the generation prompt.

On the industry side, control is the new battleground. With Pika 1.0 gaining traction and Stability AI pushing Stable Video Diffusion, raw generation quality is quickly turning into table stakes. Runway is betting that offering fine-grained artistic control is what keeps paying professionals in its ecosystem.

We think Runway’s strategy is spot-on. Prompt-based generation without spatial control was never going to work for real commercial pipelines. That said, the risk remains consistency: when you force multiple independent motions into a single shot, any model hallucination or visual artifact becomes twice as noticeable.