What is WALDO?​

A deep learning Artificial Intelligence (A.I) can detect the human behavioral characteristics of a user within a video game.

Waldo is a project, led by volunteers, that aims to detect cheaters. We plan to train an A.I. to understand how humans play video games via a visual machine learning program.

Once the program understands how humans play video games based on gameplay footage, we can then feed it gameplay footage to determine if the player in the footage is receiving assistance from a 3rd party "hack" or "cheat" program.

What is WALDO's goal?​

The first goal to achieve is the detection of the most prevalent "closet hack" which is humanized aim-assist. We will build a program that can extrapolate mouse movement data from gameplay footage via 3D imaging and optical flow. Once accurate mouse data can be logged from gameplay footage successfully, we can then begin to train the A.I. what human aim looks like in a video game. This will be done by providing data to the program from gameplay that has been verified as human. With enough data, we will have a trained A.I. program that can tell if the aim within any given piece of gameplay footage is human or machine-assisted.

How can I help?​

Do you have programming skills? A.I. / Machine learning skills? Research skills or experience in graphic design?

Check out the Projects tab for a list of the features we're currently working on and hop into the welcome discussion thread or the discord server to start contributing.

Will WALDO be opensource?​

We strive for the open-source community to help, so our main goal is to make everything open source where anyone in the future can contribute and for the openness of the project.

When will WALDO be released?​

There is no set date when WALDO will be usable. However, we want it up and running as soon as possible, and it is with your help!