Show HN: Ctrlb-decompose: Strip the noise from logs before sending to LLMs

Ctrlb-decompose is a tool that compresses raw log files into structured patterns by stripping noise and identifying key variables, reducing millions of log lines into dozens of actionable patterns with statistical analysis. The tool works as a CLI, browser-based WASM application, or Rust library, using a two-stage pipeline that extracts timestamps, encodes structured data (IPs, durations, status codes), and clusters similar logs to identify patterns with quantile statistics, anomalies, and severity scoring. This preprocessing makes logs significantly more efficient to send to LLMs by reducing complexity by over 99%.