Beeble AI Launches Canvas, a Node-Based AI Compositor for VFX Pipelines
Beeble AI has launched Canvas, a compositing platform designed to bring AI video models into professional VFX and virtual production workflows through a node-based interface. Rather than treating AI generation as a standalone step, Canvas is built around the idea of combining generative models with traditional compositing utilities and automating repetitive pipeline tasks across multiple shots simultaneously.
The platform ships with native access to two of Beeble's existing models: SwitchX, which handles video-to-video transformation, and SwitchLight, which generates physically based rendering passes including normal maps and other surface information useful in VFX compositing. An integrated AI rotoscoping tool is also included. Beyond Beeble's own models, Canvas is designed to accept external generative models as nodes in the workflow, which means studios could in principle connect it to third-party generation tools they already use.
The node-based approach is familiar to anyone who has worked in Nuke, Fusion, or similar compositing software, which lowers the conceptual barrier for VFX artists evaluating whether Canvas fits their existing practices. The batch iteration capability - applying a pipeline to a set of shots in sequence - addresses one of the more tedious aspects of integrating AI generation into production work, where re-running variations manually across many clips becomes a significant time cost.
Beeble is primarily known for its relighting and video-to-VFX tools used in smaller production and virtual production contexts. Canvas represents a broader push toward becoming a platform rather than a collection of individual tools, though how it performs on large-scale professional productions will depend on factors like render speed, model quality at scale, and integration with existing asset management systems.
