VSCode extension that monitors your whole Mac (CPU / memory / disk), uses AI to turn the numbers into checkable cleanup actions (uninstall an extension, clear a cache, kill a process), and executes the ones you pick — safely.
Small Macs (8 GB RAM) stutter under VSCode / Trae, and the culprit roams across the whole machine — CPU one moment, memory pressure or a bloated extension cache the next. Raw numbers don't tell you what to do. This extension watches the whole machine, and an AI assistant translates the snapshot into concrete, executable cleanup steps you can tick and run in one click.
- Whole-machine dashboard: live CPU / memory / disk, plus top processes — in an editor-side panel
- AI cleanup suggestions: send the snapshot to your model (z.ai / GLM via the Anthropic-compatible endpoint, or any OpenAI-compatible endpoint), get back a checkable list of actions with reasons, risk levels, and ready-to-run commands
- Tick and run, safely: pick the suggestions you trust, confirm, and the extension executes them. Every command passes a strict allowlist (uninstall extension / clear a specific
globalStoragecache / kill a pid);rm -rf /, pipe injection, and out-of-scope paths are refused - Adjustable thresholds & alert cooldown: tune CPU / memory / disk thresholds and the alert interval right in the panel, with suggested values and ranges
- Two entry points: the status-bar alert's "View details" and the Activity Bar icon both open the panel
- macOS (uses built-in
top/ps/df/memory_pressure— nosudo, no Xcode Command Line Tools)
npm install
npm run build # or npm run watch for continuous build
# In VSCode, press F5 to open the Extension Development HostPackage as .vsix and install:
npm install -g @vscode/vsce
npm run package
# In VSCode: "Extensions → Install from VSIX", pick dist/*.vsix| Command | Description |
|---|---|
| Resource Monitor: Open Panel | Open the resource dashboard |
| Resource Monitor: Start / Stop Monitoring | Start / stop polling |
| Resource Monitor: Diagnose (AI cleanup) | Collect a snapshot and ask the AI for cleanup suggestions |
| Resource Monitor: Set API Key | Store your AI API key (encrypted via SecretStorage) |
| Option | Default | Description |
|---|---|---|
resourceMonitor.interval |
5 |
Polling interval (seconds) |
resourceMonitor.alertCooldown |
300 |
Min seconds between alerts on the same metric (panel-adjustable) |
resourceMonitor.threshold.cpuTotal |
75 |
Whole-machine CPU % that triggers an alert (above) |
resourceMonitor.threshold.cpuProcess |
80 |
Single-process CPU % that triggers an alert (above) |
resourceMonitor.threshold.memoryUsed |
80 |
Memory usage % that triggers an alert (above) |
resourceMonitor.threshold.diskUsed |
85 |
Disk usage % that triggers an alert (above) |
resourceMonitor.ai.baseUrl |
https://api.z.ai/api/anthropic |
AI endpoint base URL |
resourceMonitor.ai.model |
glm-5.2 |
AI model id |
resourceMonitor.ai.protocol |
anthropic |
anthropic or openai |
- Collectors (
src/collectors/) read CPU / memory / disk / VSCode processes via shell commands — zero runtime dependencies src/ai/calls your model directly (Node 18fetch); the API key lives inSecretStorage, never in settings or logssrc/cleaner.tsis the safety gate: a command allowlist that refuses anything dangerous before execution
Copyright (c) 2026 All Contributors. Licensed under the MIT License.
Attribution: If this project helps you, a ⭐ on GitHub and retaining the copyright notice are appreciated. In derived works, please credit "Resource Monitor (https://github.com/xhqing/ResourceMonitor)".
Citing this project:
Resource Monitor — VSCode extension for whole-machine resource monitoring and AI-assisted cleanup.
https://github.com/xhqing/ResourceMonitor