High-performance GOAP AI with Burst-compiled planning, visual ScriptableObject authoring, smart sensors, and confidence-based AI memory.GOAP Hub is a complete, high-performance Goal-Oriented Action Planning framework for Unity 6.3+. Build intelligent AI agents entirely through ScriptableObject-driven visual authoring.Core Features:Burst-Compiled A Planner* — Plans are computed off the main thread using a 128-bit bitmask world-state representation with Hamming distance heuristic. Zero allocations at runtime.128-Bit Tag System — World state is represented as a compact 128-bit bitmask where each bit maps to a belief. Supports up to 128 unique beliefs per brain with lightning-fast bitwise precondition checks, effect application, and state comparison — all Burst-compatible.Signal System — A typed, decoupled messaging framework that lets actions communicate with GameObjects. Define custom signals, attach providers to agents and listeners to targets, and wire it all up through ScriptableObject strategies. Signals resolve targets from the Blackboard, report handled/unhandled results, and integrate cleanly into the action pipeline.Visual Authoring Pipeline — Define Brains, Goals, Actions, Beliefs, and States as ScriptableObjects. Drag, drop, and configure AI behavior without writing a single line of code.Recollection System — Confidence-based AI memory with temporal decay. Agents remember stimuli, investigate last-known positions, and react based on confidence thresholds.Pluggable Sensors — Optical, Auditory, and Sphere sensors with staggered update scheduling for performance-friendly perception.Neural Wiring — Slot-to-Key mapping enables reusable action assets across different agent types without duplication.Utility Curves — AnimationCurve-based modifiers on Goals and Actions allow dynamic priority and cost adjustments driven by Blackboard values.Type-Safe Blackboard — Zero-boxing generic storage with editor debug overrides for rapid iteration.GOAP Hub Editor — Full visual editor suite with Library panel, Planner Graph, Neural Circuit view, and live play-mode debugging that follows your agent step-by-step.Claude Code Ready — Ships with CLAUDE.md project files so developers using Claude Code can immediately understand the framework's architecture, conventions, and APIs — making AI-assisted development with GOAP Forge seamless from the start.Genre-Agnostic — Suitable for stealth, strategy, simulation, RPG, FPS, or any genre requiring intelligent agent behavior. Fully customizable and extensible through code when needed.Includes a ready-to-play Guard Demo scene showcasing patrol, chase, and investigation behaviors with optical sensing and Recollection.Requires: Unity 6.3+, Burst 1.8.27+, Collections 2.6.4+, Mathematics 1.3.3+Features:Burst-compiled A* planner with zero-allocation runtime planning128-bit bitmask world-state representation (TagMask128) with Hamming distance heuristicScriptableObject-driven authoring — Brains, Goals, Actions, Beliefs, States, Keys, and SlotsNeural Wiring system with Slot-to-Key mapping for reusable action assets across agent typesTyped Signal system for decoupled agent-to-GameObject messagingConfidence-based Recollection system with temporal memory decayType-safe Blackboard with zero-boxing generic storage and editor debug overridesUtility Curves (AnimationCurve) for dynamic Goal priority and Action costPluggable sensors: Optical, Auditory, and Sphere with staggered update schedulingGOAP Hub editor window: Library panel, Planner Graph, Neural Circuit view, and live play-mode debuggerPost-Process system for chaining behaviors on action strategies (e.g., NavMesh movement + wait)Includes Guard Demo scene with patrol, chase, and investigation behaviorsSupported OS:Windows, macOS, Linux (Unity Editor)All Unity build targets supported at runtimeLink to documentation:https://rogueguardianstudios.com/products/goap/Requirements:Unity 6.3 (6000.3) or newerBurst 1.8.27+Collections 2.6.4+Mathematics 1.3.3+GOAP Hub was developed as a collaboration between the developer and Claude (Anthropic) via Claude Code. AI was used throughout the development process as a coding partner — assisting with implementation, refactoring, debugging, editor tooling, and documentation writing (including this listing).All architecture decisions, system design, API surface, and feature direction were driven by the developer. Every line of code was reviewed, tested, and validated by a human before inclusion. The included CLAUDE.md files are a byproduct of this workflow — they document the framework's architecture and conventions, and are shipped with the asset so other developers using AI coding tools can benefit from the same context.
