System Control Panel

Turn your NVIDIA GPU into a self-hosted AI workstation with real-time monitoring, a semantic knowledge base, and full service management — all from a single dashboard.

What is this?

A local-first control panel for Windows machines with NVIDIA GPUs. It combines:

🖥️ System Monitoring — CPU, RAM, GPU temp, VRAM, fan speed, power draw
🧠 Brain KB — Semantic knowledge base with GPU-accelerated embeddings (bge-m3)
🔧 Service Management — Start/stop Ollama, Qdrant, and AI services from the UI
🏥 Health Diagnostics — Detect memory leaks, handle leaks, and zombie processes
🎮 Mode Switching — One-click gaming/work mode (frees VRAM, manages power plans)
🌐 Network Monitor — Multi-hop latency tracking with traceroute, jitter, and MOS scoring

Everything runs locally on your machine. No cloud. No subscriptions.

Architecture

┌──────────────────────────────────────────────────────────────────────┐
│                         YOUR WINDOWS HOST                            │
│                                                                      │
│   ┌──────────────────────────────┐   ┌──────────────────────────┐   │
│   │  Dashboard API (port 9001)   │   │  NVIDIA GPU              │   │
│   │  FastAPI + static frontend   │   │  VRAM-accelerated AI     │   │
│   │  ── Routers ──────────────── │   │  bge-m3 embeddings       │   │
│   │  /api/system    /api/gpu     │   │  Ollama models           │   │
│   │  /api/services  /api/health  │   └──────────────────────────┘   │
│   │  /api/brain     /api/network │                                   │
│   │  /api/bluetooth /api/mode    │   ┌──────────────────────────┐   │
│   └──────────────────────────────┘   │  Qdrant (port 6333)      │   │
│                                       │  Vector database         │   │
│   ┌──────────────────────────────┐   └──────────────────────────┘   │
│   │  Ollama (port 11434)         │                                   │
│   │  Local LLM server            │   ┌──────────────────────────┐   │
│   │  Embeddings + reranking      │   │  Brain CLI               │   │
│   └──────────────────────────────┘   │  Ingest / query / list   │   │
│                                       └──────────────────────────┘   │
└──────────────────────────────────────────────────────────────────────┘
              ▲                                    ▲
              │ LAN (auto-detected IP)             │ localhost
              │                                    │
    ┌─────────┴───────────┐              ┌─────────┴───────────┐
    │  Browser / curl      │              │  brain_cli.py        │
    │  Any LAN device      │              │  Antigravity / IDE   │
    └─────────────────────┘              └─────────────────────┘

Default Ports

Service	Port	Description
Dashboard	9001	Web UI + REST API
Qdrant	6333	Vector database
Ollama	11434	LLM server (optional)

All ports are configurable via environment variables. See Configuration.

Quick Start

Prerequisites

Windows 10/11 with an NVIDIA GPU
Python 3.10+ with pip
NVIDIA drivers (for nvidia-smi)
Qdrant binary (download)
Ollama (optional, for LLM features) (download)

Installation

# Clone the repo
git clone https://github.com/JairFC/system-control-panel.git
cd system-control-panel

Then double-click INSTALL.cmd — that's it.

Alternatively, from PowerShell: .\setup.ps1

The installer automatically:

✅ Checks prerequisites (Python 3.10+, NVIDIA GPU)
✅ Creates virtual environment & installs PyTorch with CUDA
✅ Downloads Qdrant vector database
✅ Creates .env configuration
✅ Creates a Desktop shortcut with a custom icon
✅ Sets up auto-start with Windows (tray icon)

After installation

Double-click "System Control Panel" on your Desktop.

The tray icon appears in your taskbar — it auto-starts the dashboard and opens the browser.

Dashboard URL: http://localhost:9001

Management Commands

.\manage.ps1 start           # Start dashboard + Qdrant (background)
.\manage.ps1 start -Dev      # Start in dev mode (foreground, auto-reload)
.\manage.ps1 stop            # Stop everything
.\manage.ps1 restart         # Restart
.\manage.ps1 status          # Check what's running
.\manage.ps1 tray            # Re-launch the tray icon
.\manage.ps1 open            # Open dashboard in browser

Configuration

All settings are loaded from environment variables. Create a .env file in the project root (see .env.example):

Variable	Default	Description
`SCP_LAN_IP`	(auto-detected)	LAN IP for the dashboard
`SCP_TRUSTED_CLIENTS`	(empty)	Comma-separated allowed client IPs
`SCP_DASHBOARD_PORT`	`9001`	Dashboard API port
`SCP_OLLAMA_PORT`	`11434`	Ollama server port
`SCP_QDRANT_PORT`	`6333`	Qdrant vector DB port
`SCP_QDRANT_EXE`	`qdrant\qdrant.exe`	Path to Qdrant binary

The LAN IP is auto-detected on startup. Override only if detection fails.

API Reference

Base URL: http://localhost:9001

All endpoints return JSON.

System Stats — `GET /api/system`

Returns CPU, RAM, and top processes.

{
  "cpu": { "load": 16, "cores": 8, "threads": 16 },
  "ram": { "total_gb": 31.7, "used_gb": 23.2, "free_gb": 8.4, "percent": 73 },
  "processes": { "total": 362, "top": [...] }
}

GPU Stats — `GET /api/gpu`

Returns GPU hardware stats from nvidia-smi.

{
  "name": "NVIDIA GeForce RTX 3080",
  "vram_used_mb": 6166, "vram_total_mb": 10240, "vram_percent": 60,
  "temperature": 41, "utilization": 23,
  "power_draw": 78.5, "fan_speed": 41,
  "compute_apps": [...]
}

Service Control — `GET /api/services`

Returns status of managed services (Ollama, Qdrant, Brain embedder).

# Start a service
curl -X POST http://localhost:9001/api/services/ollama/start

# Stop a service
curl -X POST http://localhost:9001/api/services/qdrant/stop

# Restart
curl -X POST http://localhost:9001/api/services/ollama/restart

Brain KB — `/api/brain`

Semantic knowledge base with GPU-accelerated embeddings.

# Ingest a document
curl -X POST http://localhost:9001/api/brain/ingest \
  -H "Content-Type: application/json" \
  -d '{"text": "Important note", "space": "personal", "title": "My Note"}'

# Query the brain
curl -X POST http://localhost:9001/api/brain/query \
  -H "Content-Type: application/json" \
  -d '{"question": "What did I note?", "contexts": ["personal"], "top_k": 5}'

System Health — `GET /api/health`

Detects memory leaks, handle leaks, and zombie processes. Returns fixable issues with one-click remediation.

Mode Switching — `POST /api/mode/{mode}`

Mode	Effect
`clean`	Kill leaky processes, flush temp files
`gaming`	Stop AI services, free VRAM, High Performance power plan
`work`	Start AI services, Balanced power plan

Network — `GET /api/network`

Multi-hop latency monitor with traceroute discovery, jitter calculation, packet loss tracking, SLA compliance scoring, and MOS (Mean Opinion Score).

Bluetooth — `GET /api/bluetooth`

MediaTek Bluetooth service control.

Brain CLI

A command-line tool for interacting with the Brain KB:

# Check status
python brain_cli.py status

# Ingest a fact
python brain_cli.py ingest --text "Python 3.12 is the latest stable" --space coding --title "Python version"

# Query
python brain_cli.py query "What Python version is latest?" --top_k 5

# List all memories
python brain_cli.py list
python brain_cli.py list --space personal

# Batch ingest from JSON
python brain_cli.py ingest-batch facts.json

Project Structure

system-control-panel/
├── setup.ps1                  # One-click installer (run this first!)
├── manage.ps1                 # Management script (start/stop/status)
├── tray.py                    # System tray icon (optional)
├── brain_cli.py               # Brain KB CLI
├── requirements.txt           # Python dependencies
├── .env.example               # Configuration template
├── LICENSE                    # MIT License
│
├── dashboard/
│   ├── api/
│   │   ├── main.py            # FastAPI app entry point
│   │   ├── config.py          # Configuration (env vars + auto-detect)
│   │   ├── middleware.py       # IP whitelist + cache control
│   │   ├── routers/
│   │   │   ├── system.py      # CPU, RAM, processes
│   │   │   ├── gpu.py         # GPU stats (nvidia-smi)
│   │   │   ├── health.py      # System health diagnostics
│   │   │   ├── bluetooth.py   # Bluetooth service control
│   │   │   ├── services/      # Service management (hexagonal)
│   │   │   │   ├── router.py      # HTTP routes
│   │   │   │   ├── ollama.py      # Ollama control + models
│   │   │   │   ├── qdrant.py      # Qdrant control
│   │   │   │   └── embed.py       # Brain embedder + power
│   │   │   ├── network/       # Network monitor (hexagonal)
│   │   │   │   ├── router.py      # HTTP routes
│   │   │   │   ├── monitor.py     # Ping monitor engine
│   │   │   │   ├── tracert.py     # Route discovery
│   │   │   │   ├── scoring.py     # SLA / Quality / MOS
│   │   │   │   ├── alert_log.py   # JSONL alert logger
│   │   │   │   └── constants.py   # Thresholds + regex
│   │   │   └── brain/         # Knowledge base (RAG)
│   │   │       ├── router.py      # API endpoints
│   │   │       ├── embedder.py    # GPU-accelerated bge-m3
│   │   │       ├── search.py      # Hybrid search
│   │   │       ├── chunking.py    # Doc chunking
│   │   │       └── memory.py      # Conversation memory
│   │   └── utils/
│   │       └── process_utils.py   # Subprocess + HTTP helpers
│   └── frontend/
│       ├── index.html         # Dashboard SPA
│       ├── css/style.css      # Design system
│       └── js/                # Modular JS components
│
├── scp_tray.vbs               # Silent tray launcher
└── install_tray.ps1           # Startup shortcut installer

Polling Recommendations (for integrations)

Endpoint	Interval	Latency	Notes
`/api/system`	5-10s	~50ms	Lightweight (psutil)
`/api/gpu`	5-10s	~200ms	nvidia-smi subprocess
`/api/services`	10-30s	~500ms	Port checks + health probes
`/api/health`	30-60s	~400ms	PowerShell diagnostics
`/api/network`	1s	~10ms	SSE or polling (when active)

Dependencies

Python 3.10+ with CUDA-enabled PyTorch
FastAPI + Uvicorn — ASGI web framework
sentence-transformers — GPU-accelerated embeddings (bge-m3)
psutil — Cross-platform system monitoring
nvidia-smi — GPU stats (included with NVIDIA drivers)
Qdrant — Vector database for Brain KB
Ollama — Local LLM server (optional, for generation/reranking)

License

MIT — use it however you want.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

System Control Panel

What is this?

Architecture

Default Ports

Quick Start

Prerequisites

Installation

After installation

Management Commands

Configuration

API Reference

System Stats — `GET /api/system`

GPU Stats — `GET /api/gpu`

Service Control — `GET /api/services`

Brain KB — `/api/brain`

System Health — `GET /api/health`

Mode Switching — `POST /api/mode/{mode}`

Network — `GET /api/network`

Bluetooth — `GET /api/bluetooth`

Brain CLI

Project Structure

Polling Recommendations (for integrations)

Dependencies

License

About

Uh oh!

Releases

Packages

Uh oh!

Contributors

Uh oh!

Languages

Name		Name	Last commit message	Last commit date
Latest commit History 1 Commit
assets		assets
dashboard		dashboard
.env.example		.env.example
.gitignore		.gitignore
INSTALL.cmd		INSTALL.cmd
LICENSE		LICENSE
README.md		README.md
brain_cli.py		brain_cli.py
install_tray.ps1		install_tray.ps1
manage.ps1		manage.ps1
requirements.txt		requirements.txt
scp.cmd		scp.cmd
scp_tray.vbs		scp_tray.vbs
setup.ps1		setup.ps1
tray.py		tray.py

Folders and files

Latest commit

History

Repository files navigation

System Control Panel

What is this?

Architecture

Default Ports

Quick Start

Prerequisites

Installation

After installation

Management Commands

Configuration

API Reference

System Stats — GET /api/system

GPU Stats — GET /api/gpu

Service Control — GET /api/services

Brain KB — /api/brain

System Health — GET /api/health

Mode Switching — POST /api/mode/{mode}

Network — GET /api/network

Bluetooth — GET /api/bluetooth

Brain CLI

Project Structure

Polling Recommendations (for integrations)

Dependencies

License

About

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

System Stats — `GET /api/system`

GPU Stats — `GET /api/gpu`

Service Control — `GET /api/services`

Brain KB — `/api/brain`

System Health — `GET /api/health`

Mode Switching — `POST /api/mode/{mode}`

Network — `GET /api/network`

Bluetooth — `GET /api/bluetooth`

Packages