Raymon gives us alerts when predictions are wrong and helps find out why, all with minimal setup.


Get started

Get started with our open source library

$ pip install raymon

Raymon offers a standalone, open source Python library that lets you profile your data and model health.

Discover our LIBRARY on gitHUB  →
Use the Raymon Observability Hub

The Raymon observability hub helps you monitor data and model health, and makes it easy to debug your systems. It offers alerting, slice-and-dice dashboards and data inspection functionality.

The observability hub complements the open source library, is extendable and can be self hosted.

Docs Raymon
examples Raymon
Demo image data
Demo structured

At a glance ...

Extensible data & model
quality checks

Raymon helps you set up data quality and model quality checks with a few lines of code. We currently offer out-of-the-box metrics for structured and vision data, and allow you to define your own metrics for any data type!

  • Extract features describing data quality, data novelty, model confidence and prediction performance of your data.
  • Generate reports for data drift and model degradation.
Track model health

Auto-configured monitoring
and alerting

Raymon provides easy-to-set-up monitoring of all relevant metrics and alerts you when your data or model quality drops, globally or for slices of your data.

  • Benchmark different slices of your production data against each other to expose slices with reduced performance.
  • Get clear, actionable alerts when things break down so you can maintain system quality.

Track, troubleshoot & improve without hassle

Raymon allows you to easily slice-and-dice your dashboards, metrics and predictions to debug issues. Raymon also helps you inspect specific predictions and find valuable data to improve your models.

  • Log valuable data & ground truth for building high-quality datasets and improving your models
  • Analyse data, data health and model performance globally or for specific slices.