-
-
QuadStick Device Examples
-
QuadStick Device
-
QCM
Inspiration
Setting up a mouth-operated QuadStick for every new game takes time. A wrong binding can make the game unplayable.
What it does
Access Agent researches a game's controls and compares them with 131 real profiles. It proposes a setup, asks about uncertain choices, and waits for approval before writing.
How we built it
Python runs the research and bounded AI agent. C# validates every binding using the real QuadStick rules. Avalonia shows the results on a picture of the controller.
Challenges we ran into
Raw controller codes were hard to understand. We replaced them with game actions such as “Dash: soft puff” and made sure the model could never write files directly.
Accomplishments that we're proud of
The agent uses real user history, shows its sources, and leaves unanswered controls untouched. Invalid changes reject the whole write.
What we learned
AI is useful for research and uncertain decisions. Validation and file writing are safer with deterministic code.
What's next for Access Agent
Test with more QuadStick users and ensure personalized agent for each person's abilities and preferences.
Log in or sign up for Devpost to join the conversation.