BatCave Monitor

BatCave Monitor is a local-first resource cockpit for Windows and Linux. It shows the machine underneath the machine: CPU pressure, memory, disk and network movement, process triage, runtime health, and the little permission-shaped holes where the operating system says "not today."

This is a public preview. It is useful now, honest about what it cannot see, and intentionally boring about privacy: BatCave reads local telemetry and keeps it local.

What It Shows

• A live process explorer with CPU, memory, disk, and estimated per-process network columns.
• A system inspector for CPU, memory, disk, and network detail, including logical-core charts and runtime health.
• A selected-process panel with PID, parent PID, CPU, kernel CPU, memory, private bytes, disk totals, thread count, handle count, access state, process path, and network attribution when available.
• Pause/resume, manual refresh, refresh cadence, search, sort, focus modes, chart history length, and theme controls.
• Explicit quality states when data is warming up, partially available, blocked by permissions, or unavailable on the current platform.

BatCave does not pretend. If ETW, eBPF, /proc, /sys, PDH, or process permissions are unavailable, the app keeps running and marks the affected metric honestly instead of painting fake numbers over the crack.

Preview Status

BatCave is ready for source-based testing and local preview builds.

• Windows and Linux native telemetry collectors are implemented.
• The Tauri app can run as a native desktop shell or as a browser-only fixture UI for layout testing.
• Windows bundles currently produce an unsigned executable and NSIS installer.
• Linux builds produce .deb and AppImage bundles.
• Installer signing and automatic updater work are still future distribution work.

Try It

Install prerequisites first:

• Node.js 24
• A current stable Rust toolchain
• On Linux, the WebKitGTK/GTK/Tauri native packages installed by scripts/install-linux-deps.sh

From the repository root on Windows:

cd src\BatCave.App
npm install
cd ..\..
powershell -NoProfile -ExecutionPolicy Bypass -File scripts/run-dev.ps1

Run only the browser UI with deterministic fixture telemetry:

powershell -NoProfile -ExecutionPolicy Bypass -File scripts/run-dev.ps1 -WebOnly

From the repository root on Ubuntu/Debian:

bash scripts/install-linux-deps.sh
cd src/BatCave.App
npm install
cd ../..
bash scripts/run-dev.sh

Run the Linux browser-only fixture UI:

bash scripts/run-dev.sh --web-only

Validate And Build

Run the full Windows validation workflow:

powershell -NoProfile -ExecutionPolicy Bypass -File scripts/validate-tauri.ps1

Run the Linux equivalent:

bash scripts/validate-tauri.sh

For faster app-level checks from src/BatCave.App:

npm run verify
npm run tauri:dev:windows
npm run tauri:build:windows

On Linux, use:

npm run verify
npm run tauri:dev:linux
npm run tauri:build:linux

Windows release builds emit the release executable and unsigned NSIS installer under src/BatCave.App/src-tauri/target/release. Linux builds emit .deb and AppImage bundles under src/BatCave.App/src-tauri/target/release/bundle.

Privacy And Local Data

BatCave is local-only by design. Do not add outbound tracking, remote logging, hosted collection, or surprise network dependencies.

Runtime state, settings, warm cache, helper snapshots, and logs stay under:

• Windows: %LOCALAPPDATA%\BatCaveMonitor
• Linux: $XDG_DATA_HOME/BatCaveMonitor or ~/.local/share/BatCaveMonitor

Theme preference is stored in browser localStorage under batcave.monitor.theme.

Platform Notes

• Windows per-process network attribution uses ETW over the kernel TCP/IP provider. If the kernel logger cannot start or access is denied, BatCave reports the reason and continues.
• Windows admin mode can launch a local elevated helper for richer snapshots. If elevation is denied or unavailable, BatCave falls back to standard access.
• Linux aggregate telemetry uses /proc and /sys. Optional per-process network attribution uses bpftrace/eBPF when the host has the needed permissions or capabilities.
• Browser fixture mode is for UI work. It is deterministic on purpose and is not proof of native collector behavior.

Benchmarks

Run a headless runtime benchmark:

powershell -NoProfile -ExecutionPolicy Bypass -File scripts/run-benchmark.ps1 -BenchmarkHost core -Platform x64 -Ticks 120 -SleepMs 1000

Capture a reusable benchmark baseline artifact:

powershell -NoProfile -ExecutionPolicy Bypass -File scripts/capture-benchmark-baseline.ps1 -BenchmarkHost core -Platform x64

Linux equivalents are available at scripts/run-benchmark.sh and scripts/capture-benchmark-baseline.sh.

More Documentation

• App runbook covers native/browser run modes, app scripts, and platform troubleshooting.
• Runtime telemetry covers the Rust runtime store, native collectors, quality states, admin helper behavior, and benchmark surfaces.

Contributing

Keep the app local, explicit, and boringly reliable. Match the existing Rust/Tauri/Svelte boundaries, preserve the snake_case JSON contracts, and run the narrowest meaningful verification before opening a PR.