The Role of AI in Apex Engine: From Assistant to Collaborator
We have some lofty plans for Apex Engine, and though we know it will be a lot of work ahead of us, we are excited to be able to share some of the design plans for how AI can be used in Apex Engine. Though AI is not an "easy button," it is a transformative tool that, when properly integrated, can change how creators, developers, and studios build interactive experiences. That belief is shaping our AI strategy for Apex Engine.
Learning From You, Not Overriding You
AI should not interrupt your flow or take over your work. It should support how you already build and help reduce the time spent on repetitive tasks or setup. In Apex Engine, AI will act as an assistant that learns from your workflow and adapts to your style. Whether you are writing game logic, designing environments, or iterating on physics systems, the goal is to enhance your creativity, not replace it.
Leveraging LLMs and SLMs for the Right Tasks
We plan to use a combination of Large Language Models (LLMs) and Small Language Models (SLMs) to support different parts of the development pipeline. LLMs will power broader, context-heavy tasks such as code generation, system design suggestions, and help documentation. SLMs will be used for fast, local decision-making like labeling objects, suggesting physics parameters, or generating on-the-fly metadata for scene elements. This allows us to keep performance high without sacrificing smart assistance.
Why We Are Building Our Own LLM
The privacy and security of our clients and their IP is critical. That is why we are building a custom, in-house LLM tailored specifically for Apex Engine. During early planning, we explored a few potential partnerships with external AI providers. While some offered proprietary licensing, we discovered that even under those terms, we would still be required to share training data to improve or fine-tune their models.
For us, that was a clear line we would not cross.
Sharing training data, especially from live projects, simulations, or unreleased IP, poses unnecessary risks not only to our engine and technology but to your work, your team, and your clients. That simply would not do.
By developing our own model, we ensure that all training, processing, and optimization happens within our secured environment. No data is sent out, and no project is exposed to third parties. This protects your creative rights, ensures compliance with industry regulations, and gives you full control over your pipeline.
Our LLM will be purpose-built for 3D development, learning and adapting to real-time workflows across industries including gaming, simulation, digital twins, immersive learning, and enterprise applications.
Phase One: Code Assist
Our first AI integration will focus on code assistance. Developers will have access to context-aware autocomplete, intelligent script suggestions, and inline help that understands the game logic and engine architecture. Whether you are prototyping new mechanics or cleaning up existing code, the AI will be there to assist, always learning from how you write.
Phase Two: Smarter Physics and Terrain Generation
Once code assist is stable, we will expand into physics simulation tools and terrain generation. AI will help fine-tune collision settings, ragdoll behavior, joint constraints, and environmental forces. For terrain, the AI will assist in procedural generation with rulesets you can teach and modify. This will help create more natural, biome-aware environments that match your creative intent.
Phase Three: Generative Placement of Actors and Assets
Eventually, we plan to introduce generative AI placement tools that help build scenes faster. You will be able to describe an area, sketch a layout, or use templates, and the engine will assist with placing props, actors, lighting, and gameplay elements. This will be especially useful for designers working with large worlds or interactive simulations. AI will recognize design patterns and suggest improvements based on previous iterations.
Looking Ahead
We are building AI into Apex Engine as a true partner in development, not a gimmick or shortcut, but a practical assistant that evolves with your project. From smart code suggestions to terrain tools and generative world-building, our roadmap is all about giving creators more control, better speed, and less friction.
And with our custom LLM solution, you can rest assured that innovation will never come at the cost of privacy or security.
We are just getting started, but we are excited to bring AI into your workflow the right way, with purpose, flexibility, and a focus on real results.