Whistle AI
A lightweight, intelligent log monitoring tool that uses local LLMs to analyze logs in real-time, identify critical events, and notify you before they become problems.
curl -sSL https://raw.githubusercontent.com/wsmlby/whistle/refs/heads/main/install.sh | bashKey Features
Intelligent Log Analysis
Instead of static rules, Whistle AI uses a LLM to understand log context, distinguishing between routine messages and critical events to reduce alert fatigue.
Adaptive Filtering
The AI continuously and automatically maintains an ignore list of benign log patterns, improving accuracy and reducing false alarms over time.
Private & Offline-First
Integrates with HoML to run on your own hardware. Your logs and data never leave your network.
Easy Deployment
Installs as a single binary with a one-line command and automatically sets up its own systemd service for you.
Flexible Monitoring
Monitor kernel messages, specific systemd service units, or any custom log file. Whistle is designed to fit your monitoring needs.
Extensible Notifications
Get notified via Slack, email, or custom webhooks. Whistle ensures you get critical alerts wherever you are.
How It Works with HoML
Running Whistle AI with HoML gives you a powerful, private log analysis pipeline. Your logs are streamed to a local LLM running on the HoML server, which then intelligently decides if an alert needs to be sent. Configuration is dead simple. Just point Whistle AI to your HoML endpoint:
whistle config llm --base_url http://localhost:8000/v1 --model your-local-model