hyperTuned.app

Dynamic kernel sysctl telemetry, automated vnode exhaustion relief, and AI workload tuning presets for macOS.

1. Executive Summary

Heavy multi-agent concurrency and background file indexing routinely saturate macOS kernel limits. When APFS vnode tables saturate (e.g. kern.num_vnodes reaching 100% of kern.maxvnodes), the operating system suffers catastrophic micro-stutters as the kernel continuously evicts and reallocates file cache structures under severe lock contention.

hyperTuned.app provides intelligent, real-time macOS kernel supervision and dynamic tuning. Sampling kernel limits every 2 seconds, hyperTuned visualizes vnode pressure, open file descriptors, process tables, and memory compressor throughput—automatically applying authorized kernel parameter elevations (kern.maxvnodes: 2M -> 4M) to restore optimal operating system headroom without requiring system restarts.

2. Core Capabilities

• Real-Time Kernel Telemetry: Live inspection of kern.num_vnodes, kern.maxvnodes, kern.num_files, kern.maxfiles, and vm.compressor_page_count.
• Dynamic Lock Contention Relief: Eliminates APFS file system micro-stutters by dynamically doubling vnode ceilings when sustained concurrency is detected.
• One-Click Workload Presets: Instant switching between "AI Fleet Heavy Load", "Standard Developer", and "Battery Conserve" operational profiles.
• Persistent Boot Management: Automates persistence across macOS restarts via /Library/LaunchDaemons and /etc/sysctl.conf.