In Flick, a 3D studio frames the shots before generation

The Flick platform integrates the 3D Stage space to set up cameras and characters before generation. A free alternative to exports from Blender.

Video models know how to create a convincing image, but are much less capable of obeying a directorial intent. Flick, a platform dedicated to AI-generated film, addresses this with 3D Stage, an integrated 3D space where you prepare your shot before handing it over to its models. Until now, obtaining this level of control required modeling in Blender, exporting, and then re-importing the render as a reference.

The scene is built from an image or a simple description, then populated from a library of assets where you can pull characters, cameras, and props. Each element can be moved, rotated, and resized using the mouse or by entering exact values, making it possible to match multiple shots or maintain the same scale from one scene to another.

Animation can be adjusted to several levels of precision. Ready-to-use actions can be dragged onto a character or their timeline. A movement path can be drawn by placing points on the ground, using Bézier curves to smooth the trajectory and a choice of pace, from a slow walk to a sprint. A game mode assigns actions to keys, lets you control the character in real time, and records the performance, which returns to the timeline as a continuous animation. More experienced users can manipulate the rig, joint by joint, using keyframes.

The camera has its own settings, including focal length, aspect ratio, and handheld shake to break up the stillness, plus two movement modes: target tracking that follows a character, and a rail mode where it glides between two points like a dolly shot. A rehearsal allows you to check framing and timing before recording, with the take then returning to the Flick canvas for editing or generation. The publisher presents the tool as free and claims a quality comparable to what is achieved in Blender.