Relight footage, build masks and combine generation with node-based compositing in Beeble’s filmmaking workspace.
Beeble AI is a filmmaking workspace for editing and compositing footage with AI. It accepts recorded or generated material and provides tools for relighting, background removal, pass generation and other finishing tasks. The product is aimed at artists who need control over an existing shot, not only a prompt that produces an unrelated new clip.
SwitchX provides mask-guided relighting and replacement. Background removal produces alpha mattes, while the VFX Pass Generator supplies information such as depth and physically based rendering passes for further work. Other tools include 3D Relight, SDR-to-HDR processing and upscaling. Beeble emphasises preserving the incoming frame format and using explicit masks to protect selected regions, which gives artists a way to define what should remain unchanged.
The platform supports PNG and EXR outputs, including EXR sequences for HDR workflows. These formats matter when a shot needs to enter a professional compositing pipeline rather than end as a compressed preview. The result should still be reviewed over the whole sequence for edges, temporal consistency and the intended lighting effect.
Canvas is a node-based environment that combines generation and compositing. Artists can connect Beeble’s native tools with supported external models, branch a workflow, compare variations and reuse a graph. Templates offer a starting point for tasks such as relighting, sky replacement or set extension. The graph keeps the steps visible, allowing a change to one stage without rebuilding the whole creative process.
Individual subscriptions provide monthly credits, and higher tiers expand allowances and selected upload capabilities. Team arrangements add pooled credits and administration, while enterprise offerings provide additional identity and support options. The official plans advertise commercial usage rights, subject to the applicable terms and rights in the input. Teams should verify model availability and credit costs for the chosen operation before processing a long sequence. A trial shot with representative motion and edges is a better assessment of workflow fit than assuming every shot will match a marketing demonstration.
