XRCore is a modular framework for building AI-driven assistants in XR applications, combining perception, event-driven logic, and agent reasoning systems.XRCore is render pipeline agnostic.The framework focuses on runtime logic, event systems, and AI agent orchestration, and does not rely on specific rendering features.It works with Built-in Render Pipeline, URP, and HDRP without modification.The included demo scene uses only basic Unity primitives and standard lighting to ensure compatibility across all pipelines.XRCore is a modular AI agent framework for Unity XR designed to accelerate the development of intelligent assistants, guided interactions, and context-aware experiences.It provides an event-driven architecture that connects perception systems, task orchestration, agent reasoning, and behaviour execution into a unified runtime pipeline. Developers can build XR applications where virtual agents detect the environment, interpret user actions, and respond with real-time instructions, guidance, or automated behaviours.The framework separates perception, reasoning, and interaction into independent layers, allowing developers to integrate different detection providers, AI models, or decision-making strategies without modifying the overall system architecture.XRCore includes pluggable perception providers, the XRCoreEventBus for decoupled communication, XRGuideAgent for contextual decision-making, and XRTaskRunner for structured task progression.The package also includes a demo scene showcasing gaze-based interaction, contextual agent guidance, and an automatic Setup Wizard that quickly configures the required components in the scene for fast integration.XRCore is ideal for XR assistants, training simulations, industrial guidance systems, interactive onboarding, and spatial computing applications.Core Features• Modular architecture for building AI-driven assistants in Unity XR applications• Event-driven system connecting perception, interaction, and agent decision-making• XRGuideAgent for contextual instructions, guidance, and intelligent behaviour execution• XRTaskRunner for task orchestration, step progression, and structured workflow management• Pluggable perception providers through the IXRDetectionProvider interface• Included detection providers for simulation, raycast-based interaction, and Sentis-ready integration• Agent reasoning interfaces supporting rule-based logic, state machines, and AI-driven decision systems• Setup Wizard for automatic scene configuration and fast project integration• Demo scene showcasing gaze-based interaction, object detection, and contextual agent guidance• Compatible with Built-in Render Pipeline, URP, and HDRP without additional configurationTechnical Highlights• Decoupled architecture powered by XRCoreEventBus for flexible system communication• Modular agent reasoning pipeline with interchangeable reasoning strategies• Easily extensible detection provider system for custom perception workflows• Compatible with Unity Sentis and external AI inference services• Designed for scalable XR assistants, training systems, and context-aware applications• Asset Store-ready package structure with clean separation between runtime, editor tools, and demo contentAI-assisted tools were used during the development process to support coding, documentation drafting, and architectural exploration.The final implementation, architecture design, testing, and integration were performed and validated manually to ensure correctness, stability, and compliance with Unity development guidelines.
