Pathfinder Patch

Anyone here remembering that weird path patch half year ago? (Yeah absolutely correct, this one we had to revert a patch later because it broke everything.) We’re more than happy to tell you that, finally, we managed to release these features in a very sophisticated and highly usable manner. On top of this, we not only added functionality for quickly constructing and following paths. We also managed to fully integrate pathfinding right into our steering approach. For now, we do this utilizing Unity’s inbuilt NavMesh technology, but in the future, we plan to release adapters for the most commonly used pathfinding plugins out there.

Good things come to those who wait (and buy the Pro version, of course)! Let your agents taking advantage of asynchronous pathfinding with the new inbuilt components without losing the flexibility and natural expressiveness of context steering. Polarith AI now combines the best of both worlds in one plugin: Pathfinding for mastering complex level structures and context steering for local avoidance, formations and much more.




As usual, here are the release notes for patch version 1.4:


  • [Pro] Added AIMFollowWaypoints behaviour to follow a path or an arbitrary collection of points
  • [Pro] Added AIMLinearPath for quickly creating your own paths or patrols
  • [Pro] Added AIMUnityPathfinding, an adapter for Unity’s NavMesh system so that AIMFollowWaypoints can utilize this information
  • [Pro] Extended overall API such that the AIMFollowWaypoints behaviour can be easily used with your custom Pathfinding solution


  • Improved custom editors, they now follow Unity’s standard with respect to multi-selection and prefab emphasizing


  • Fixed a bug which broke the multi-selection capabilities of custom editors (if you implemented your own editor based on AIMBehaviourEditor, you need to override the BehaviourProperty)
  • Fixed a bug where Unity crashed on calling Polarith AI API methods in a specific order (AIMContext.ResizeObjectives, thanks @Simon for pointing this out)
  • Fixed a bug in AIMContext which caused the indicator gizmo to flicker based on the set UpdateFrequency
  • Fixed the PlaneGizmo such that it is consistent with our other gizmos


What’s Next?

We’re not done yet! We’ve got so many ideas we want to implement for you. Currently, we’re busy developing our long-term feature for full 3D mesh-based sensors. Moreover, we’ll take our time to make things shine which are already there. That said, in our current development branch, we just improved the performance of massive agent setups by a factor of two. Of course, we will also polish usability here and there, but the most existing feature for the next patch will be a completely overhauled layer system without breaking backward compatibility. Combining behaviours will be much easier and much more intuitive. The creation of queues, boids, flocks and herds will take you no more than a few minutes. For that, we are currently preparing some decent video tutorials, too. Stay tuned and let us know any bugs! Thank you.