AI Hardware Monitor
A fullscreen Rust terminal dashboard for the machine running your models — CPU, GPU, NPU, Coral TPU, memory, disk and network — with an AI-process filter and six colour themes.
A live terminal dashboard for the hardware behind an AI workload: aggregate and per-core CPU, RAM, swap and temperature; Intel/AMD/NVIDIA GPU, Linux NPU and USB/PCIe Coral TPU detection with whatever driver telemetry is exposed; disk and network throughput; and Linux pressure-stall figures, all as braille history charts and gauges across five keyboard-driven workspaces. Process view sorts by CPU, RAM or I/O and can filter down to likely AI workloads. Unavailable metrics are shown as N/A rather than a fabricated zero — shared GPU memory is not invented as dedicated VRAM, and an open accelerator handle is not treated as proof of active inference. Six themes, Nerd Font icons with a plain-text fallback, JSON snapshot output, and AMD64/ARM64 Debian packages built in CI.