caiovicentino / apple-silicon-internals

What is apple-silicon-internals?

This Objective-C toolkit reverse engineers Apple Silicon internals, uncovering private APIs for Apple Intelligence models, Metal 4 ML pipelines, and 1009 SoC telemetry channels without sudo. Developers get CLI tools to map hardware specs, monitor real-time power across subsystems, scan frameworks for undocumented classes, and x-ray apps for entitlements and AI usage. Built for macOS betas on M-series chips, it solves the black-box problem of Apple's hardware by delivering runtime introspection and benchmarks via simple commands like `make all` followed by `probes/soc_power` or `tools/framework_scanner`.

Why is it gaining traction?

Unlike sudo-dependent tools like asitop (which shows ~20 metrics), this exposes 1009 IOReport channels for granular CPU/GPU/ANE power and thermals, plus 55+ on-device AI models and Metal tensor APIs ahead of official docs. The 16 Claude Code skills—like `/silicon-power` for energy profiling or `/silicon-deepxray` for app audits—hook AI tinkerers with instant hardware access. It's a github reverse engineering goldmine for github reverse engineering ki and reverse engineering app workflows, blending github reverse commits insights with practical Metal 4 speedups up to 2.7x.

Who should use this?

Apple platform devs optimizing ML inference on M4/M5, hardware hackers probing SoC telemetry for custom drivers, and security researchers auditing apps via entitlement dumps and framework scans. Ideal for reverse engineering informatik pros building github reverse proxy nginx alternatives or reverse engineering 3d scan tools, especially those hitting llama.cpp limits on Apple Silicon.

Verdict

Grab it for bleeding-edge Apple reverse engineering wiki knowledge if you're on macOS betas—MIT licensed with solid README benchmarks—but expect breakage on updates given 10 stars and 0.699999988079071% credibility score. Early but actionable for niche power users; skip for production.

(198 words)