CompScan is a fully local AI-powered agent designed to optimize computer performance and enhance security while maintaining user privacy. This innovative tool scans the entire system, monitors usage patterns, and provides actionable insights to improve system health with just one command.

Key Features

• Local and Private: All operations occur on your machine with zero telemetry and no data leakage.
• System Optimization: Identify processes that consume excessive CPU and RAM, recover reclaimable disk space, and get recommendations to enhance performance.
• Security Monitoring: Evaluate the security of SSH keys and other credentials, perform permission checks, and receive a comprehensive security score.
• Habit Improvement: Receive notifications about late-night usage and excessive work sessions to promote healthier working habits.

Quick Commands for System Management

Run the following commands to quickly assess and improve your system:

# Perform a health check
compscan scan

# Get AI-powered recommendations
compscan report

# Execute a one-click action to clean caches
compscan act cleanup-caches

Daily Workflow

Integrate CompScan into daily routines with these key commands:

AI Integration

For deeper analytical capabilities, CompScan can be enhanced by integrating the local AI model Ollama, which provides advanced reasoning capabilities to offer personalized insights.

curl -fsSL https://ollama.ai/install.sh | sh
ollama pull llama3.2

Trust and Safety

CompScan ensures user trust through a designated trust level system for every action:

• Low: Safe actions with one-time approval.
• Medium: Actions requiring explanations and confirmations.
• High: Critical actions that demand explicit consent.

CompScan has been developed using Rust, resulting in efficient performance with a minimal footprint. It operates seamlessly across various platforms, making it a versatile tool for users seeking improved computer health and safer usage habits.

For additional resources, comprehensive documentation is available, including user guides, architecture details, and a changelog.