Your PM2 Autopilot - Modern interactive CLI for PM2 process management with AI assistance, built with TypeScript, ESM, and pnpm.
🎬 See PM2 Pilot in Action: Watch how PM2 Pilot intelligently analyzes process errors, provides AI-powered diagnosis, and suggests actionable solutions—all through natural language interaction!
- Smart Log Analysis - Automatic error detection and categorization
- Intelligent Diagnosis - AI-powered root cause analysis with actionable suggestions
- Natural Language Interface - Ask questions in any language, get instant help
- Proactive Error Alerts - Automatically detects and explains critical issues
- Multi-Provider AI Support - Works with OpenAI GPT and Google Gemini
- Interactive Shell Interface - Slash commands and natural language
- Real-time Process Monitoring - Live metrics and performance data
- Smart Log Streaming - Enhanced log viewing with error analysis
- Colorized Output - Beautiful, readable terminal interface
- Auto-Recovery System - Process health monitoring with automatic restart
- Safe Configuration Management - Enterprise-grade config handling with file locking and atomic writes
- TypeScript - Full type safety and modern development
- ESM Modules - Modern JavaScript with pnpm package management
- Comprehensive Testing - Unit tests and BDD scenarios with Cucumber
- Fast Performance - Optimized for quick process operations
- Node.js >= 22.0.0
- PM2 installed globally (
npm install -g pm2) - pnpm package manager
Install PM2 Pilot globally to use it anywhere:
npm install -g pm2-pilotAfter installation, start the interactive CLI:
pm2pilotOnce you start the CLI with pnpm start, you'll enter an interactive shell where you can use these commands:
/statusor/ps- Show PM2 process status/listor/ls- List all PM2 processes/restart <name>- Restart a specific process/stop <name>- Stop a specific process/start <name>- Start a specific process/all <operation>- Batch operations on all processes
/healthor/check- Run comprehensive health checks/watch <name> [mem] [cpu]- Auto-recovery monitoring with thresholds/watch stop- Stop watching processes
/metricsor/m- System metrics and health score/history- Performance history and trends/history collect- Start collecting metrics/history clear- Clear history data
/logs <name>- Stream real-time logs with automatic error detection/logs smart <name>- AI-powered log analysis with error diagnosis/grep <pattern> [name]- Search across all logs/errorsor/err- Show recent errors only
/save [filename]- Save PM2 ecosystem configuration/load [filename]- Load saved configuration
/ai <question>- Ask AI assistant about processes/ai config- Show AI configuration status/ai provider <openai|gemini>- Switch AI provider/ai providers- List available AI providers/ai presets- Show speed/quality presets for current provider/ai preset <name>- Apply speed preset (lightning, fast, smart, reasoning)/diagnose [process]- AI-powered issue diagnosis/doctor logs [process]- Comprehensive log analysis and error diagnosis/doctor errors [process]- Focused error analysis with critical issue detection/optimize [process]- Get AI optimization suggestions
/helpor/h- Show available commands/clearor/cls- Clear the terminal screen/history- View and search command history/exit,/quit, or/q- Exit the CLI
# Start the interactive CLI
pm2pilot
# Basic monitoring:
> /status
> /metrics
> /health
# Intelligent log analysis:
> /logs smart my-app # AI-powered error analysis
> /doctor logs # Comprehensive diagnosis
> /doctor errors my-app # Critical error focus
# Advanced features:
> /watch my-app 500 80 # Auto-restart if >500MB or >80% CPU
> /grep "error" my-app # Search logs for errors
> /all restart # Restart all processes
# AI-powered insights (requires OpenAI or Gemini API key):
> my app is slow, can you help? # Natural language
> /ai why is my-app slow? # Direct AI query
> /diagnose my-app # Process diagnosis
> /optimize # Optimization suggestions
# Configuration:
> /save production-config
> /load production-config
> /exitThe AI features support both OpenAI and Google Gemini. Choose your preferred provider:
# Option 1: Environment variable
export OPENAI_API_KEY=your-openai-api-key
# Option 2: Configure via CLI
> /ai provider openai
> /ai config# Option 1: Environment variable
export GEMINI_API_KEY=your-gemini-api-key
# Option 2: Configure via CLI
> /ai provider gemini
> /ai config# List available providers
> /ai providers
# Switch between providers
> /ai provider gemini
> /ai provider openai
# Check current configuration
> /ai config
# Speed presets for optimal performance
> /ai presets # Show available speed presets
> /ai preset lightning # GPT-5 nano - Ultra fast responses
> /ai preset fast # GPT-4o - Balanced speed and quality (recommended)
> /ai preset smart # GPT-5 - Highest intelligence for complex analysis
> /ai preset reasoning # GPT-5 mini - Deep thinking for complex problemsPM2 Pilot uses a unified configuration file in your home directory with safe concurrent access:
- Location:
~/.pm2pilot-config.json - Features:
- Thread-safe configuration: Uses file locking to prevent corruption
- Atomic writes: Prevents race conditions during config updates
- Auto-recovery: Rebuilds corrupted config files automatically
- Contents:
- AI provider selection (OpenAI/Gemini)
- API keys for each provider
- Model presets and active selections
- User preferences (auto-execute mode, verbosity, confirmation level)
- Provider-specific configurations
This single file manages all PM2 Pilot settings with enterprise-grade reliability and concurrent access protection.
# View current configuration
> /ai config
# The config file structure:
{
"provider": "openai",
"userPreferences": {
"autoExecute": false,
"verbosity": "detailed",
"confirmationLevel": "destructive"
},
"openai": {
"apiKey": "sk-...",
"model": "gpt-4o",
"activePreset": "fast"
},
"gemini": {
"apiKey": "AI...",
"model": "gemini-2.5-flash"
}
}PM2 Pilot is evolving toward an AI-native experience where natural language becomes the primary interface:
> my app is slow
❌ Commands must start with /. Type /help for available commands.
> /ai my app is slow
🤖 AI Assistant is thinking...
[AI response without actions]> my app is slow
🤖 Analyzing your processes...
I can see that "my-app" is using 85% CPU and 420MB memory. This could be causing slowness.
Let me check the logs for errors:
[Automatically shows recent errors]
Suggestions:
• Restart the process: restart my-app
• Check memory usage: /metrics my-app
• Would you like me to restart it now?
> yes restart it
🔄 Restarting my-app...
✅ my-app restarted successfully
📊 New metrics: CPU 12%, Memory 89MB - much better!# Process troubleshooting
> why did my-app crash?
🤖 Analyzing crash logs... Found memory limit exceeded at 2:34 AM
🔧 Suggesting: increase max_memory_restart to 800MB
# Batch operations
> restart all errored processes
🤖 Found 3 errored processes: api-server, worker-1, notifications
🔄 Restarting all... Done! All processes now online.
# Performance optimization
> optimize my cluster setup
🤖 Analysis: Your 4 instances are underutilized (avg 15% CPU)
💡 Recommendation: Reduce to 2 instances to save 180MB memory
🚀 Would you like me to scale down automatically?
# Learning and discovery
> how do I monitor memory usage over time?
🤖 I can help! Use `/history collect` to start tracking, then `/history` to view trends.
📊 Starting collection now... You'll have data in 30 seconds.- Zero Learning Curve: Start with natural language immediately
- Contextual Actions: AI automatically includes current system state
- Proactive Suggestions: Get optimization recommendations without asking
- Intelligent Automation: AI executes common tasks with confirmation
- Conversational Workflow: Build solutions through natural dialogue
- AI-Powered Assistant - Ask natural language questions about your processes
- Smart Diagnosis - AI analyzes process issues and suggests solutions
- Optimization AI - Get personalized performance improvement recommendations
- Auto-Recovery System - Automatically restart processes on memory/CPU thresholds
- Health Checks - Comprehensive health scoring with detailed diagnostics
- Performance History - Track and analyze metrics over time with trends
- Smart Batch Operations - Safe bulk operations with confirmation
- Log Search Engine - Grep across all process logs instantly
- Error Aggregation - View all errors from all processes in one place
- Safe Configuration Management - Enterprise-grade config handling with file locking and atomic writes
- Smoother UX - Interactive shell eliminates repetitive
pm2typing - Colored Output - Beautiful, readable terminal output
- Unified Dashboard - All metrics in one clean view
- Smart Alerts - Proactive notifications for issues
- Session Persistence - Stay in context while monitoring
- Graceful error handling
- Auto-recovery from crashes
- Performance trend analysis
- Configuration backup/restore
- Health monitoring dashboard
Contributions are welcome! Please see the Contributing Guide for details on:
- Development setup and workflow
- Code style and testing requirements
- Pull request process
- Project architecture and structure
For technical details, see:
MIT License - see LICENSE file for details.