Welcome to GPT-LogSpotLight, a unified suite of applications built on Python and Streamlit, harnessing the capabilities of OpenAI's language models for specialized log analysis. This umbrella solution encapsulates a variety of tools, each designed for a unique log type or use case, transforming raw log data into actionable insights. With GPT-LogSpotLight, users can interactively dialogue with the logs, extracting valuable insights about activities on their tenant.
GPT-LogSpotLight is a comprehensive platform hosting a multitude of applications, each honed to analyze a distinct type of log. These applications are built on the robust foundation of OpenAI's language models, including the current GPT-3.5-turbo-16k and future iterations like GPT-4. The objective of GPT-LogSpotLight is to offer a versatile and scalable solution for log analysis, capable of evolving alongside advancements in AI technology.
The inaugural application under this suite is LogSpectra-SecurityEvents. This Python-based application leverages Microsoft's Graph API to fetch security event logs and employs OpenAI's GPT-3.5-turbo-16k model for analysis, based on user queries. It offers a user-friendly interface for seamless interaction with the data and extraction of valuable insights.
- LogSpotlight-SecurityEvents: A Python application that fetches event logs from Microsoft's Graph API and utilizes OpenAI's GPT-3.5-turbo-16k model for user query-based analysis.
To utilize an application within the GPT-LogSpotLight suite, please refer to the ReadMe.md file within each application's directory for specific instructions. The applications are built on Python and require various libraries (check the provided requirements.txt file in each application's directory).
GPT-LogSpotLight is envisioned to expand with more applications, each tailored to a unique log type or use case. Planned enhancements include integration with Azure OpenAI Service and Office 365 Management API, and the introduction of a data pre-processing step to accommodate larger alert volumes. The suite is also designed to adapt to future OpenAI models, ensuring its continued relevance and effectiveness.
This project is licensed under the terms of the MIT license.