We are officially pushing the Video Remix Agents into Open Beta.
This suite unlocks a massive end-to-end video remixing use case: you can now take raw brand assets and UGC, and automatically remix them to generate 10x the performance-optimized social spots.
What you need to know:
The feature consists of a three-stage pipeline:
Video Splitter Agent: Intelligently breaks down raw video into targeted scenes and generates a frame-accurate clip manifest. (Note: This is also incredibly powerful as a standalone tool if you just need rapid scene discovery from a long shoot!).
Video Remix Script Agent: Takes that manifest and automatically writes structured storyboards. You can apply a video template for strict branding, or let the AI assign scenes intelligently.
Video Stitcher Agent: Assembles the storyboard, clips, and text overlays into a cohesive, rendered final video.
How to test it (The Workflow Way):
While you can explore these agents interactively in Chat (up to 14 videos in one go), the real magic happens in Workflows for scaled production.
Navigate to Workflows and create a new canvas
Connect the nodes in order: Video Splitter
➡️Video Remix Script ➡️Video Stitcher.
Feed a large collection of raw videos into the Splitter and watch the automated data flow generate final stitched spots at scale.
Q: When you have a remixed video output is it possible to take that in to the editor to make fine adjustments? I.e. Does that remixed video show up as a file with separate scenes that you can adjust individually?
