gonzo

Why we included this project

Gonzo reads log streams straight from the command line, pulling from stdin, local files, Kubernetes, OTLP receivers, and CloudWatch. It renders what it gets as severity heatmaps and word-frequency stats you can filter with regex or by attribute, in a terminal UI modeled on k9s, so you can investigate an incident without bouncing between a slow dashboard and a raw tail. The optional AI layer is the part that stands out: point it at OpenAI, Anthropic Claude, Ollama, LM Studio, or any OpenAI-compatible endpoint and it flags unusual patterns and suggests likely root causes on the live stream. Everything runs locally and works fully offline with local models, which suits engineers who want AI-assisted debugging without shipping log contents to a third party. Teams already piping kubectl or OpenTelemetry output will find it a sharper signal than plain tail.