Blaze AI Engine

Universal NPC Development in Unity

Blaze AI serves as a comprehensive engine for developers looking to implement complex non-player character (NPC) logic within the Unity ecosystem. Compatible with all versions including Unity 6, the system is designed to handle a broad spectrum of AI types, ranging from simple wanderers and zombies to sophisticated tactical units. The engine’s architecture allows for the creation of melee brutes, agile climbers, and shooters capable of utilizing cover tactics. Because the system is modular, it does not lock developers into a specific workflow, providing the flexibility to switch behaviors dynamically at runtime, such as transitioning an AI from an aggressive attack state to a fleeing state.

Streamlined Animation and Root Motion

One of the more distinct features of this engine is its approach to animation. The system bypasses the need for complex trigger and transition setups within the Unity Animator. Instead, animations are dragged into the Animator and selected via a dropdown menu within the Blaze interface. This method is designed to keep the animator controller organized and clean while still providing full support for root motion. By handling the transitions internally, the engine simplifies the process of achieving commercial-grade movement without the manual overhead typically associated with state machine management.

Advanced Navigation and Off-Mesh Links

Navigation is handled through a system that includes built-in local avoidance, ensuring that AIs do not collide with one another while moving through an environment. In companion mode, this avoidance extends to the player, allowing NPCs to maintain a natural distance. A standout technical capability is the support for off-mesh links, which enables NPCs to perform complex navigational tasks such as jumping across gaps, falling from heights, and climbing ladders. This functionality is fully compatible with both root motion and the companion system, allowing for verticality in level design that NPCs can navigate autonomously.

Combat Logic and Spare States

The engine supports various combat styles and interactions, including AI vs AI combat where NPCs can engage each other until a knockout or death occurs. For more nuanced behaviors, the system utilizes “Spare States,” which are custom states that can be triggered at any time. These are used for emotes, taunts, blocking attacks, or stun animations, allowing the AI to react to game events with specific audio or visual cues. Additionally, the system features multi-target vision, detection bars, and tactical maneuvers like strafing and retreating. NPCs can also be configured to “call for help,” alerting nearby allies to the player’s presence.

Optimization and Performance Scaling

To maintain performance across different hardware, the engine is built around a single Update() call per AI. It incorporates several optimization layers, including distance culling and configurable vision performance options to reduce the processing load of NPCs that are far from the player. Audio is handled via scriptable objects to keep the system lightweight. For developers needing deeper integration, the engine provides extensive APIs and public property access, allowing for the adjustment of AI behavior through code during runtime.

Companion Functionality

Beyond enemy AI, the engine includes a dedicated companion mode. This allows NPCs to fight alongside the player and respond to specific commands. These commands include following the player, staying at a location, moving to a specific point, or attacking a designated target. The companion system utilizes the same core navigation and animation features as the enemy AI, ensuring that friendly NPCs possess the same agility and environmental awareness as their hostile counterparts.

Practical Takeaway

Blaze AI provides a centralized solution for NPC logic that replaces fragmented systems for navigation, animation, and combat. By utilizing the off-mesh link support and the simplified animation workflow, developers can implement vertical movement and complex character interactions while maintaining a clean project structure and optimized performance overhead.