Skip to main content

Butterflybrain

Based on game build 714014 | Last updated: 2026-02-27

Overview​

The ButterflyBrain component defines the behavior tree for a butterfly entity in Don't Starve Together. It orchestrates core AI actions such as wandering toward flowers, collecting pollen, returning home (to a flower), and fleeing from threats. This brain integrates with the game's behavior system (via behaviours/) and leverages shared panic triggers from BrainCommon. It relies on the pollinator component to determine pollen collection status and the skilltreeupdater component to check for player-specific modifiers that affect threat perception.

Dependencies & Tags​

  • Components used:
    • pollinator: Checked via self.inst.components.pollinator:HasCollectedEnough() to determine if the butterfly should return home.
    • skilltreeupdater: Checked via guy.components.skilltreeupdater:IsActivated("wormwood_bugs") in the RUN_AWAY_PARAMS filter to suppress panic in players who have unlocked that skill.
  • Tags:
    • Uses {"scarytoprey"} for threat detection (in RunAway).
    • Uses {"flower"} via GetClosestInstWithTag for flower seeking.
  • Behaviors referenced:
    • RunAway, Wander, DoAction, FindFlower
    • Shared triggers: BrainCommon.PanicTrigger, BrainCommon.ElectricFencePanicTrigger

Properties​

No public properties are initialized in the constructor. The class uses local constants (RUN_AWAY_DIST, STOP_RUN_AWAY_DIST, etc.) and a behavior tree (self.bt) stored on the instance.

Main Functions​

ButterflyBrain:OnStart()​

  • Description: Initializes and sets the behavior tree for the butterfly. This method is called when the brain takes control of the entity. It builds a priority-based behavior tree that evaluates actions in order: panic responses (if triggered), threat avoidance, flight-to-home logic, flower seeking, and finally general wandering.
  • Parameters: None.
  • Returns: None.

Events & Listeners​

None identified. The brain does not register or dispatch events directly. It operates solely via its behavior tree evaluation loop.