Kraken for Unreal Update

October 5, 2016

We’ve been working on various integration approaches for Fabric and Unreal, but we hadn’t been able to solve the problem of deploying Fabric content as part of an Unreal .exe until recently. This meant that Fabric for Unreal was limited to live on-set type scenarios rather than shipping packaged content. Alongside the general integration with Unreal, we have also been working on getting Kraken rigs working in Unreal. With the new deployment capability we are now delighted to say we have a work flow for authoring Kraken rigs in Maya/Softimage (and soon, 3DS Max) and deploying those rigs as part of an Unreal .exe (with no middleware).

We are now able to use the Kraken Rigging Framework to build a character rig in Maya, then build the same exact rig as a code-only version within Unreal (and potentially any other engine). Essentially it’s a KL extension which provides a data structure with the exact same features as the rig in Maya. This code-only rig can be used in a range of scenarios:

  • inside Canvas for processing a pose based on arbitrary animation data.
  • used together with our Blocks feature to compute poses on crowds – by simply using an array of these rigs.
  • use the code-only rig inside of Unreal Engine with a new plugin we are working on. This allows you to inject your actual animation data (not the baked forward kinematics animation) into a real-time engine.

By ‘animation data’ we mean: Animation Curves, Alembic caches or USD files – but also of course live animation sources such as retargeted motion capture streams etc. Essentially: Anything you can build in KL and Canvas!

If you would like to join the beta, email: with the following:
– your name
– your company
– how much time you have available for testing per week
– your technical proficiency (Fabric requires some technical chops)

We aren’t giving access to the Kraken for Unreal work for another month or two, but if you want to get familiar with Kraken in the meantime, you can get more information here: