Skip to main content

Anomstack

Painless open source anomaly detection for your metrics! 📈📉🚀

🚀 Easy Setup & Deploy

Get anomaly detection running in minutes with Docker, GitHub Codespaces, Dagster Cloud, or local Python. No complex infrastructure required - just bring your metrics and we'll handle the ML.

🌟 Flexible Data Sources

Connect to BigQuery, Snowflake, ClickHouse, DuckDB, SQLite, MotherDuck, Turso, and more. Define metrics with SQL or custom Python functions. Works with your existing data stack.

🧠 Intelligent Detection

Powered by PyOD machine learning algorithms with configurable thresholds. Plus optional AI-powered anomaly detection using LLM agents for more nuanced analysis and explanations.

📧 Smart Alerts

Get notified via Email and Slack with beautiful ASCII art plots and detailed context. Includes snoozing, feedback system, and daily summaries to reduce alert fatigue.

📊 Beautiful Dashboard

Monitor your metrics with a modern FastHTML + MonsterUI dashboard. View trends, anomaly scores, feedback ratings, and manage alerts all in one place.

⚙️ Highly Configurable

Customize everything from ML algorithms to alert routing with flexible YAML configs. Override settings per metric batch and add custom preprocessing functions.

Supported Data Sources & Platforms

🗄️ Data Sources

PythonBigQuerySnowflakeClickHouseDuckDBSQLiteMotherDuckTursoRedshift 🚧

☁️ Model Storage

LocalGoogle Cloud StorageAWS S3Azure Blob 🚧

Deploy Anywhere

Choose the deployment option that works best for you

🐳 Docker

Self-hosted with Docker Compose

☁️ Dagster Cloud

Serverless managed deployment

🚀 Codespaces

GitHub cloud development

🔧 Replit

Browser-based deployment

🐍 Python Env

Local virtual environment

Ready to detect anomalies in your metrics?

Join the growing community of developers using Anomstack for reliable, open-source anomaly detection.

⭐ Star us on GitHub if you find Anomstack useful!