Today, we are excited to announce the launch of the public beta for Universal Game AI (UGA) — a platform that allows anyone to train their own AI agents for virtually any computer game.

UGA was built with a single goal in mind: making modern game AI accessible to everyone. There is no longer a need to spend months learning machine learning, writing complex code, or building training infrastructure from scratch. Everything you need is already integrated into a single, unified system.

After installing the application on your computer, you can create a new project, configure training settings, record a dataset directly from gameplay, and train an agent based on your own playstyle.

Learn Through Gameplay

The core idea behind Universal Game AI is simple: you play, the agent learns.

While recording a dataset, the application captures your actions, keyboard inputs, mouse movements, and other gameplay information required for training. Once recording is complete, you can start the training process and create your own AI agent capable of reproducing the learned behavior and performing tasks autonomously within the game.

This approach makes game AI training understandable and accessible even for users with no technical or machine learning background.

Community and Project Sharing

Universal Game AI is more than just a desktop application — it is a complete ecosystem built around sharing and collaboration.

By connecting the application to the online platform, you gain access to a growing community where users can publish their own projects and explore solutions created by others.

A project may contain one or multiple trained agents, configuration files, descriptions, and additional resources. This makes it easy to distribute complete solutions and allows other users to start using them with minimal setup.

Powerful Tools in One Place

UGA brings together every major step of the game agent workflow into a single platform:

• project creation and management;  
• dataset recording directly from live gameplay;  
• agent training through an intuitive graphical interface;  
• testing and running trained models;  
• project import and export functionality;  
• integration with an online platform for sharing projects;  
• centralized management of agents and datasets.

This Is Just the Beginning

The current release marks the first public beta version of Universal Game AI. Many new features, improvements, and tools are already planned and will be introduced in future updates.

Right now, you have the opportunity to be among the very first users to explore the platform, influence its development, and help shape the future of Universal Game AI.

More information about the platform and its capabilities can be found in the documentation.

Download Universal Game AI today, test its features, share your feedback, and become part of a community building the next generation of game AI technology.