More Than Just an AI Model

The AIRPGHUB platform utilizes the latest advances in AI to provide the best gameplay experience possible to our users. While most other platforms provide a direct interaction with an AI model, AIRPGHUB provides an enhanced experience that requires no additional knowledge or setup by the user.

Our system uses several key technologies to provide this expanded experience:

Agentic Frameworks

Our system has specialized AI agents for completing certain tasks, these agents are built through a combination of model selection, prompt engineering, tool calling, structured outputs, and other techniques. Most chat applications require users to build their own agents or choose one from a selection, additionally communication between these agents can require manual steps or programming by the user.

Our system, not only has these agents built-in but also the inter-communication between them and processing of their outputs automatically handled. This removes the typical emphasis on prompt engineering and more careful choice of language when interacting with AI systems.

Inference Time Compute

As the agentic framework implies, we do not ask an AI for one single response, but query (ask) the AI many times when generating a single response to the user. As such, it may take a bit longer to receive a response from our platform, but the quality of the response is much higher.

If you are familiar with OpenAI's O1 model, and how it spends "more time thinking before responding" (https://openai.com/o1/), our system is built on a similar premise.

Background Processing

Additionally, to help alleviate the load of the inference time compute, we have a background processing system that allows us to perform tasks in the background while the user is interacting with the platform. This allows our AI GM master to continue working, while the player is reading the response or engaging with other generated content. While playing, you will notice that the right-hand side of the screen will be actively receiving and rendering updates, which reflect some of these background tasks.