FLORA introduces Action Node

FLORA launches Action Node, a feature bringing 21 non-AI editing tools to its canvas to eliminate back-and-forth with external post-production software.

FLORA, a node-based creative environment dedicated to generative AI workflows, integrates a new type of node called Action Node, which brings traditional editing operations directly onto the canvas, without resorting to external editing software. The team's stated goal: to eliminate back-and-forth between generation tools and post-production pipelines.

More than 21 actions are accessible from the "+" menu, under Utilities. For images: color grade, filter, tint, blur, rotation, flip, aspect ratio change, and duplication. For videos: splitting by scenes or into equal parts (Split Video), color grading, stitching with transitions (Stitch Videos), frame extraction (Extract Video Frames), playback reversal, speed up or slow down, boomerang, green screen removal, conversion of a still image into an animated shot using the Ken Burns effect, or transformation of a video into a frame grid. For text: split, find & replace, concatenation.

These actions are deterministic and non-AI, guaranteeing identical output for the same input, at a reduced cost and execution time compared to calling a generative model. They chain together with existing generation nodes, allowing for seamless alternation between creative steps and formatting operations. Combined with the Batch Node, they open up serial use cases, for example, to apply the same color grading to multiple variations of an image.

The demonstration shared by the team chains together the generation of a handbag, application of filters and blur, then assembly into a video sequence, all taking place without leaving the canvas. The feature is immediately accessible at app.flora.ai.