Give footage a shared frame.
The reconstruction pipeline prepares video, estimates camera positions, and packages a Gaussian-splat scene for the browser.
An interactive workspace for connecting footage, spatial reconstruction, and narrative. Ariadne brings source cameras, annotations, and a story timeline into an explorable 3D scene.

Footage records what one camera could see. Understanding a situation means connecting that viewpoint to other perspectives, a place, and a moment in time.
Ariadne is the spatial reconstruction layer of LabyrinthOS. The project combines video reconstruction with an authoring workspace where cameras, annotations, paths, and story beats share the same scene.
The reconstruction pipeline prepares video, estimates camera positions, and packages a Gaussian-splat scene for the browser.
Inspect source cameras, switch scene layers, and return to saved views. Spatial context stays close to the material being examined.
Place annotations, draw paths, and arrange story beats on a timeline. A scene becomes a guided sequence that someone else can follow.
The interface distinguishes observed information, estimates, and manually authored elements. A camera position derived from alignment carries a different meaning from an observation in the original footage.
That distinction shapes the authoring experience. Adding a marker or a path helps explain a scene, but does not make the addition source evidence.

Ariadne brings together interactive graphics, reconstruction workflows, and the craft of guiding attention. The interesting design problem is making a complex scene readable without losing the context behind it.
This is collaborative work within the Heraklion ecosystem and remains in active development. These captures show the current studio with its bundled demonstration content.
Local app captures, September 2026. Butterfly scene: Spark sample asset. Demo annotations and camera positions are illustrative; these images do not document a reconstruction from original footage.
Have a complex product to bring into focus?