Black Ops II recompiled into a native ARM64 Mac app, built with Claude Opus 5.5
bo2-mac converts your own Steam copy of Call of Duty: Black Ops II into a native ARM64 macOS app that runs the campaign, Zombies and bot multiplayer without Wine, Rosetta or an emulator.
bo2-mac takes the three Black Ops II executables from your own Steam copy and translates their 32-bit x86 code into ARM64 machine code ahead of time, on your Mac. The repository contains no game files. According to the README, the code is lifted to LLVM IR with Remill, rewritten by the project's own passes, and linked with a hand-written Win32 runtime whose Direct3D 11 is drawn by Metal. The author reports the campaign at 51 to 60 frames per second on an Apple M1, Zombies at 57 to 60, and multiplayer against bots at 53 to 60.
The author writes that they used Claude Opus 5.5 to build the tool. The README says AI agents wrote the shims, bisected millions of calls per run and read tens of thousands of lines. The human set the goal, corrected results against what appeared on screen, and rejected flattering conclusions. The README calls this one game, months of work and heavy manual intervention. Known problems include heat, which the author says is still to be fixed, and a conversion that takes about 1 hour 45 minutes on an M1.
Unfinished items include split-screen, which the author says is coming soon, and LAN mode, which is on the roadmap. The README also says the multiplayer and Zombies match rules were rewritten in a clean room. The reason given is that Steam no longer ships the original daily patch file.
In the author's words
So I used Claude Opus 5.5 to create a tool that converts your copy of the game (from Steam) to a native ARM64 Mac app, with almost 60fps in all areas.
From the README or the author's post. We checked that the sentence is there, not that it is true.
What to take from it
A small interpreter that runs the original bytes can serve as a fallback for untranslated code, trading speed for stability instead of crashing. The README's habit of writing predictions down before measuring is also worth copying in any performance project.
Built with
See it
AI wrote this entry from the project's own README or its author's post. That AI built the project is the author's statement, quoted here. Nobody at IKANGAI ran or tested it. How this list is made