We taught ClickStack to read your logs like a detective novel
Blog post from ClickHouse
HyperDX has introduced a new feature, AI Summarize, which transforms complex log entries, traces, and patterns into narrative summaries to bridge the gap between raw telemetry and human understanding, making them more comprehensible and engaging. This feature uses hand-written phrase pools and data extraction to create themed narratives that adapt to the context, such as Detective Noir for errors and Shakespearean Drama for performance issues, without relying on external large language models (LLMs), ensuring zero cost, zero data sharing, and zero latency. The humorous and creative narratives, like portraying a swift Redis call as a peregrine falcon or a persistent log pattern as a forgotten chorus, aim to make technical data more relatable and entertaining while maintaining accuracy with real OpenTelemetry attributes. This innovation is designed to remain client-side, ensuring privacy and transparency, and although it currently does not use AI, the infrastructure for potential LLM integration is in place, with feedback invited for future development.