Join our channel

Follow 3DCGHUB on Telegram for new resource drops, migration updates, and community notices.

Join Telegram

Safety & quality

Resources are manually reviewed before listing to improve quality and reduce obvious risks.

No account required
Artificial Intelli

Utility AI - Wise Feline Ultimate V3 (NPC Memory, Emotions, Smart Objects, GAS)

Explore Utility AI - Wise Feline Ultimate V3 for Unreal Engine, featuring dynamic utility scoring, Remembrance memory, GAS, and Smart Object integration.

Categories: Unreal Engine, Artificial IntelligencePublished Nov 22, 2026
DownloadFree Download

Get this resource

Compatibility

  • Unreal Engine 5.1 - 5.6
  • Asset Version1.7.0 - 2.1.0
Format
.7z
File size
7.9 MB

Ready to download

Resource archiveWise Feline.7z
View details

Technical specifications

CategoryArtificial Intelligence
Engine & PlatformUnreal Engine 5.1 - 5.6
Asset Version1.7.0 - 2.1.0
Package Format.7z
File Size7.9 MB
Download AccessFree Download
PublishedNov 22, 2026
Resource overview

Dynamic Decision-Making with Wise Feline Ultimate

A villager working in the fields often faces competing demands. Finishing a daily task takes priority over a slight appetite, but as hunger grows, lunchtime eventually overrules field work. If a wolf suddenly appears on the path, both labor and meal plans need to be abandoned immediately. Wise Feline Ultimate gives Unreal Engine non-player characters a structured way to handle these shifting priorities. Instead of navigating rigid, tangled state trees, the character assigns a score to each candidate action based on its immediate perception of the world, determining whether an active task should continue or yield to an interruption.

Decision-making within this framework operates on continuous utility evaluation. Every action an agent can take is vetted against the current state of the game environment. When circumstances change—whether through rising danger, shifting survival stats, or secondary opportunities—the utility framework re-evaluates the validity of current tasks against emerging options, allowing NPCs to step away from repetitive scripts and respond fluidly to ambient events.

Considerations, Curves, and Action Scoring

At the center of the utility calculation are considerations. A consideration gathers raw data from the game—such as an actor's hunger level, spatial distance to an adversary, or remaining ammunition—and normalizes it. A response curve then converts that raw measurement into a definitive numerical score. By adjusting the shape of the mathematical curve, designers dictate how an NPC perceives urgency. A linear curve yields a steady rise in motivation, while an exponential curve keeps an NPC passive until a threshold is crossed, triggering sudden, decisive action.

Developers can assemble actions and considerations using either Blueprint or C++, combining them directly inside the behavior editor. The framework includes 17 editable presets that serve as baselines for common behaviors. Actions are modular assets that can be shared across multiple behavior definitions, and full behaviors can be dynamically added or removed from an agent at runtime. When picking what to do next, the agent can either select the absolute highest-scoring action for strict, deterministic choices or sample from the top candidates using uniform or score-weighted randomness to introduce organic variability into group behavior.

Deep Diagnostics with the Decision Debugger and Memory Inspector

Complex utility scoring can become opaque without robust diagnostic tools. Wise Feline Ultimate addresses this by letting developers pin an individual agent brain inside a dedicated decision debugger. While the game runs, the tool provides real-time visibility into which actions were weighed, which action maintained execution priority, and how each input value, response curve, and target score contributed to the outcome. When diagnosing rapid logic changes, the display can be frozen to inspect a frame without interrupting the background recording. Furthermore, decision data surrounding an action switch can be exported for post-run analysis even after a Play In Editor session has terminated.

Diagnostic support extends into Unreal Engine's native profiling suite. The system includes built-in hooks for both the Visual Logger and the Gameplay Debugger. For projects tracking social relationships and emotional states, the Memory and Emotion Inspector exposes the internal states managed by the Remembrance module, ensuring developers can verify why an agent holds a grudge, displays fear, or warms up to friendly players.

Living Village Simulation and Emotional Remembrance

Included with the Ultimate package is Wise Feline Remembrance, a subsystem dedicated to tracking personal history and emotional sentiment. A past encounter can directly color an agent's future choices through four specialized considerations:

  • The general mood of the NPC across time.
  • The specific emotional disposition held toward a distinct actor.
  • The elapsed game time since a particular event occurred.
  • The total count of matching events stored in the character's memory.

These mechanics are showcased in the Living Village sample, built as a standalone Blueprint demonstration in Unreal Engine 5.8. The sample models seven individual villagers as they balance hunger, physical fatigue, socialization, and vocational tasks. If an NPC witnesses a player stealing goods from a market stall, that witness can seek out a town guard to report the crime. Returning the stolen parcel prompts the merchant to adjust transaction pricing, yet the independent witness continues to remember the original theft. The sample demonstrates the combined power of Ultimate decision-making, spatial Influence Maps, and integration toolsets working simultaneously within an interconnected community.

Integrating Smart Objects, GAS, and Behavior Trees

Utility architecture does not require discarding conventional Unreal Engine AI workflows. An entire existing Behavior Tree can be encapsulated and run as a single utility action. The system can read from and write to standard AI Blackboards, run and evaluate Environment Query System (EQS) queries, and command agents to navigate toward the resulting query positions.

Through the separate, free WFIntegrations package, utility logic connects directly to Epic's modern gameplay systems. Actions can claim, navigate to, and interact with Smart Objects placed throughout a level. For titles utilizing the Gameplay Ability System (GAS), the integration allows utility actions to trigger specific Gameplay Abilities while utilizing specialized considerations that evaluate gameplay tags, check attribute values, or determine whether an ability is off cooldown and ready for activation.

AI Assistant Workflow and Behavior Authoring in UE 5.8

To accelerate setup and balancing, the package bundles an AI assistant skill tailored for utility behavior design. The skill equips external language models with domain-specific knowledge regarding response curve design, spatial target filtering, and action lifecycle handling. Designers can query the assistant to draft specialized curves—such as balancing cautious reload tendencies against offensive pressure—or to troubleshoot logic anomalies where a sentry fails to break a patrol path when enemies enter range.

Setup is managed directly within the editor via Tools > Wise Feline > Getting Started, which installs bundled skills for development environments including Codex, Claude Code, and Gemini CLI, or configures prompts for web interfaces such as ChatGPT, Claude Cowork, and Gemini Gems. In Unreal Engine 5.8, developers can connect an assistant directly to the editor using the Model Context Protocol (MCP), enabling external agents to directly inspect, create, and refine behavior assets, considerations, actions, and evaluation curves inside the project workspace.

More From The Same Workflow