YouTube Content ID identifies copyrighted content by matching video and audio fingerprints. If you use gameplay footage that a publisher has claimed, Content ID will flag your video. The remix layer in the machine defeats this at the file level before upload.

How Fingerprinting Works

Content ID takes a mathematical hash of a video's pixel values and audio waveforms at regular intervals. If enough intervals match a registered asset, the video is flagged. The hack is simple: change enough pixel and audio waveform values that the match score falls below the detection threshold.

Content ID protection
Every Video Gets a Unique Fingerprint Before Upload
Crop, color shift, and speed variation are applied automatically by FFmpeg. Runs on every video in every channel.
Get the Machine --

What the Remix Layer Does

For each video the script applies a randomised combination of: a 2 to 5 percent crop, a slight hue and saturation shift, and a 0.97x to 1.03x playback speed change. Each combination is unique per video. The resulting file has a distinct fingerprint that does not match the source footage's registered hash. The visual difference to a human viewer is imperceptible.