AI Log Analyzer

Modern IT systems generate massive amounts of log data every day. These logs often contain critical information about system errors or cyber attacks, but identifying relevant patterns manually is extremely time-consuming.

The goal of this project was to automate log analysis using artificial intelligence. A local AI system evaluates log entries, detects anomalies, and prioritizes suspicious events to support administrators and security analysts.

The application was built as a full-stack system. The backend, implemented in Laravel, handles log parsing, chunking, and communication with the AI layer. A Vue 3 dashboard visualizes results through charts, filters, and anomaly tables in real time.

For the AI component, Flowise and Ollama were used to run local large language models (LLMs), ensuring data privacy and independence from external cloud services. The system follows a two-phase analysis approach, combining fast pre-processing with deeper semantic evaluation.

The entire project was developed in a Docker-based environment and deployed on a Linux server using Nginx. Particular focus was placed on performance, stability, and usability, with a target of processing around 10,000 log lines within two minutes.

During development, challenges such as API timeouts and inconsistent AI responses were addressed through retry mechanisms and optimized processing pipelines. Additionally, features such as PDF export and interactive dashboards were implemented to improve usability.

This project provided valuable insights into real-world software development, combining cybersecurity, artificial intelligence, and full-stack engineering in a practical environment.