Skip to main content

OWL Watch

OWL Watch is an AI-powered log monitoring tool that automatically investigates errors in your application logs and suggests fixes by analyzing your source code.

What It Does​

  1. Monitors log files - Watches your application logs in real-time (tail -f style)
  2. Detects errors - Identifies ERROR, WARN, FATAL messages and stack traces
  3. Investigates with AI - Uses a local LLM to analyze errors and read your source code
  4. Suggests fixes - Provides actionable code changes with explanations

Key Features​

FeatureDescription
Real-time DetectionMonitors logs continuously, captures multi-line stack traces
AI InvestigationLLM reads your code, searches for related files, traces the error
Project ProfilesUnderstands your project structure, framework, and dependencies
Web DashboardLive error stream, searchable history, investigation interface
Feedback SystemTrack investigation accuracy with Accurate/Partial/Inaccurate ratings
Webhook AlertsSend notifications to Slack, Discord, or custom endpoints

How It Works​

Log Error → Detect → Create Investigation → User clicks Investigate
         → LLM analyzes error + reads source code
         → Suggests fix with explanation
         → User applies fix and provides feedback
  1. OWL Watch detects an error in your logs
  2. Error appears in the dashboard with type and timestamp
  3. You click "Investigate" to start AI analysis
  4. The LLM examines the error and searches your codebase
  5. You receive a fix suggestion with the exact file and code to change

Supported Frameworks​

OWL Watch has built-in error patterns for common frameworks:

  • Java: Spring Boot, Spring
  • Python: Django, Flask, FastAPI
  • JavaScript: Express, React, Vue, Angular, Next.js

It works with any project - the LLM can read and understand any codebase.

Requirements​

  • Python 3.10+
  • Ollama running locally

Quick Example​

# Create a profile for your project
owl-watch study ./my-spring-app

# Start monitoring logs
owl-watch ./my-spring-app/logs/app.log --project my-spring-app

Open http://localhost:8080 to see the dashboard:

Dashboard Stats

When an error occurs, click to investigate and get a fix:

Investigation Complete

Next Steps​