Mac mini Revival: From Dust Collector to 24/7 AI Assistant with OpenClaw

After upgrading my equipment early this year, my old M2 Mac mini sat gathering dust in the corner for months. Until recently when OpenClaw went viral, I suddenly realized: this 24/7 running, performance-excessive little box could become a development powerhouse. Today I'm sharing my transformation experience.

Background: The Mac mini Dilemma

Many developers have similar experiences:

• Bought a Mac mini as a dev machine, later switched to MacBook Pro for mobility
• Put it at the office as a CI/CD server, but cloud servers are more convenient
• Want to sell it second-hand, but Mac retains value well, yet the price is unsatisfactory

M2 Mac mini characteristics:

• ✅ Powerful performance (8-core M2, 16GB RAM)
• ✅ Low power consumption (average 20W, ~$1.5/month electricity)
• ✅ Silent (virtually no fan noise)
• ❌ But no monitor, keyboard, gathering dust at home/office

Turning Point: The Emergence of OpenClaw

In January 2026, OpenClaw (formerly Clawdbot) suddenly went viral, with GitHub stars breaking 6 million within 72 hours. What is it?

OpenClaw is a self-hosted AI assistant that can:

• Connect to local file system (read/write files, execute commands)
• Integrate external APIs (GitHub, Notion, databases, etc.)
• Interact anytime via Telegram/WhatsApp/Discord
• Based on Claude API, but runs locally

The key point: OpenClaw needs a 24/7 running machine, and Mac mini fits perfectly!

My Transformation: 3-in-1 Productivity Tool

I transformed the Mac mini into:

1. OpenClaw AI Assistant (interact anytime via Telegram)
2. Remote Development Server (via VS Code Remote-SSH)
3. Production Environment Server (running client projects' live services)

Architecture Diagram

MacBook Pro (Mobile Dev Machine)
    │
    │ Tailscale (Zero-config VPN)
    ▼
Mac mini (24/7 Running)
├── OpenClaw (AI Assistant)
│   ├── Connected to Claude API
│   ├── Access local files
│   └── Telegram Bot interaction
├── Development Environment (Docker)
│   ├── MySQL
│   ├── Redis
│   └── Node.js projects
└── Production Services (PM2)
    ├── API Server
    ├── Web Application
    └── Scheduled Tasks

Implementation Steps (2 Hours Total)

Step 1: Install OpenClaw (10 minutes)

OpenClaw provides Homebrew installation:

# 1. Install OpenClaw
brew install openclawd

# 2. Configure
openclawd onboard

# Follow prompts to enter:
# - Claude API Key (from Anthropic website)
# - Telegram Bot Token (create from @BotFather)

Configuration file example (~/.openclaw/config.json):

{
  "anthropic_api_key": "sk-ant-xxx",
  "telegram": {
    "bot_token": "123456:ABC-xxx",
    "chat_id": "Your Telegram User ID"
  },
  "workspace": "/Users/ekt/code",
  "allowed_paths": [
    "/Users/ekt/code",
    "/Users/ekt/Documents"
  ],
  "tools": {
    "file_system": true,
    "terminal": true,
    "browser": false
  }
}

Start OpenClaw:

# Foreground (for testing)
openclawd start

# Background (recommended)
brew services start openclawd

# View logs
tail -f ~/Library/Logs/openclawd.log

Step 2: Configure Telegram Bot (5 minutes)

In Telegram:

1. Find @BotFather, create new Bot
2. Get Bot Token
3. Get your Chat ID (via @userinfobot)
4. Fill in OpenClaw configuration

Test:

You: "List package.json dependencies for current project"
OpenClaw: [Auto-reads file] "Project uses Next.js 14, React 18..."

You: "Create a README.md with project introduction"
OpenClaw: [Auto-writes file] "Created README.md"

You: "Run npm test"
OpenClaw: [Executes command] "Tests passed, 12 test cases"

Step 3: Configure Remote Development Environment (30 minutes)

3.1 Install Tailscale (Zero-config VPN)

# On Mac mini
brew install --cask tailscale
# Launch and login

# Also install on MacBook Pro
brew install --cask tailscale

Tailscale advantages:

• No router configuration needed
• Automatic NAT traversal
• Cross-platform support (Mac/Win/Linux/iOS/Android)
• Free tier supports 100 devices

3.2 Configure SSH Quick Access

On MacBook Pro, edit ~/.ssh/config:

Host macmini
  HostName 100.xxx.xxx.xxx  # Mac mini's Tailscale IP
  User ekt
  ControlMaster auto
  ControlPath ~/.ssh/%r@%h:%p
  ControlPersist 10m
  ForwardAgent yes

Test connection:

ssh macmini
# After success, press Ctrl+D to exit

3.3 Configure VS Code Remote-SSH

1. Install extension: ms-vscode-remote.remote-ssh
2. Press ⇧⌘P → Remote-SSH: Connect to Host → select macmini
3. Open project folder (path on Mac mini)

Experience:

• Code editing latency <30ms (feels like local)
• Terminal, debugger all run remotely
• MacBook Pro fan doesn't spin, battery life extended 50%

Step 4: Deploy Production Services (30 minutes)

Use PM2 to manage production services:

# Install PM2 on Mac mini
npm install -g pm2

# Configure ecosystem.config.js
module.exports = {
  apps: [
    {
      name: 'api-server',
      cwd: '/Users/ekt/code/production/api',
      script: 'npm',
      args: 'start',
      env: {
        NODE_ENV: 'production',
        PORT: 3001
      },
      instances: 2,  // Utilize multi-core
      exec_mode: 'cluster',
      autorestart: true,
      watch: false,
    }
  ]
}

# Start services
pm2 start ecosystem.config.js

# Auto-start on boot
pm2 startup
pm2 save

# Monitor
pm2 monit

Step 5: Configure Monitoring & Alerts (20 minutes)

Use Uptime Kuma to monitor all services:

# Run with Docker on Mac mini
docker run -d \
  --name uptime-kuma \
  -p 3001:3001 \
  -v ~/uptime-kuma:/app/data \
  louislam/uptime-kuma:latest

# Visit http://macmini-ip:3001
# Add monitors: API health checks, database, Redis, etc.
# Configure alerts: Telegram Bot notifications

Real Usage Scenarios

Scenario 1: AI Programming Assistant Anytime, Anywhere

When out and about:

Me (via Telegram): "Check production logs for errors"
OpenClaw: [SSH to Mac mini] [Check PM2 logs] "No errors in past hour"

Me: "Yesterday's new feature, users report slow login, analyze it"
OpenClaw: [Reads code] [Checks DB queries] "Found N+1 query issue in login API, suggest..."

Scenario 2: Mobile Remote Development

At a coffee shop:

# 1. MacBook Pro connects to Tailscale
# 2. Open VS Code Remote-SSH
# 3. Connect to Mac mini
# 4. Code, compile, test all run remotely
# 5. Local only handles display, battery life extended 2-3 hours

Scenario 3: Stable Production Environment

Mac mini runs production services 24/7:

• API server (PM2 cluster mode, utilizing multi-core)
• Database (MySQL Docker container)
• Scheduled tasks (node-cron hourly data collection)
• Monitoring & alerts (Uptime Kuma)

Cost:

• Electricity: ~$1.5/month
• Network: $0 (Tailscale free)
• Compared to cloud server: AWS t3.small ~$15/month

Advanced OpenClaw Usage

1. Automated Operations

Me: "Send server statistics every morning at 9 AM"
OpenClaw: [Configures cron job] "Set up, starts tomorrow"

Me: "If API response time >2s, auto-restart service and notify me"
OpenClaw: [Sets monitoring script] "Alert rule configured"

2. Code Review Assistant

Me: "Review my recent Git commits"
OpenClaw: [Reads git log] [Analyzes code changes] "Found 3 potential issues:..."

3. Documentation Generation

Me: "Generate API docs based on current project"
OpenClaw: [Scans code] [Generates Markdown] "Created docs/API.md"

Cost & Benefits

One-time Investment

Monthly Cost

Comparison with Cloud Servers

Mac mini solution is 5-10x more cost-effective!

FAQ

Q1: Is Mac mini performance sufficient?

Absolutely!

My configuration: M2 8-core 16GB

Actual load:

• OpenClaw: CPU 5%, Memory 500MB
• Production services (3 Node.js processes): CPU 15%, Memory 4GB
• MySQL + Redis: CPU 7%, Memory 1.5GB
• VS Code Remote: CPU 5%, Memory 1GB

Total: CPU 30%, Memory 8GB, plenty of headroom

Q2: What about security?

Multiple safeguards:

1. Tailscale encrypted tunnel: End-to-end encrypted, not via public internet
2. OpenClaw permission control: Only allows access to specified directories
3. SSH key authentication: Password login disabled
4. Firewall: Only necessary ports opened

Q3: What if network disconnects?

Contingency plan:

• Tailscale auto-reconnects (recovers within 5 seconds)
• PM2 auto-restarts services
• MacBook Pro can temporarily run production services (same Docker Compose config)

Q4: Will OpenClaw leak code?

No!

• OpenClaw only runs locally, code never leaves Mac mini
• Only instructions you send and Claude API responses are transmitted
• Can configure allowed_paths to limit access scope

Summary

Transforming an idle Mac mini into a OpenClaw AI assistant + remote development + production server 3-in-1 tool brings:

1. AI Programming Assistant: Interact via Telegram anywhere, helps write code, check logs, analyze issues
2. Remote Development: Seamlessly switch between coffee shop, home, office, MacBook Pro battery life extended 50%
3. Production Server: $1.5/month cost (electricity), saves $13.5/month vs cloud servers
4. Learn New Tech: Experience AI assistants, remote development, DevOps, and other cutting-edge technologies

ROI: If you already own a Mac mini, transformation cost is nearly $0, immediately usable.

If you also have an idle Mac mini, give this solution a try!

About the Author: ekent, Technical Lead at ek Studio, focusing on development tools and productivity improvements.

Related Resources:

• OpenClaw Official
• Tailscale Official
• VS Code Remote-SSH Docs