Laya vs Jev runs the classic offline T-Rex Runner game twice at once: one run driven by a local decision model over MLX, the other by a hosted Jev call, both choosing when to jump from the same game state.
Live metrics track each run's decisions as the game plays, and every run can be recorded and replayed afterward for a closer look at where the two diverged.
It is a small, self-contained way to see the speed and behaviour difference between a local and a hosted decision model on an identical task, rather than reading numbers in a table.
What you get
- Full source code under the Apache 2.0 license
- The dual T-Rex Runner harness, live metrics, and replay recording





