A Python library that uses LLMs to spot anomalies in time series data. Feed it a DataFrame, it'll tell you what's weird and why – with a verification step to filter out the false alarms. Built on LangGraph for workflow management, works across multiple columns, and you can customize the prompts for your specific domain (sales spikes vs. sensor failures vs. whatever weird data you're dealing with).
Anomaly Agent
LLM-powered anomaly detection for time series data using LangGraph