NannyML is an open-source Python library for monitoring machine-learning models after deployment. It estimates model performance without access to ground-truth labels, detects data drift and concept drift, and links drift alerts to changes in performance. NannyML Cloud builds on the library and adds alerting and webhooks. It can run as a managed application inside your own Azure or AWS subscription, or as a software-as-a-service option on NannyML's infrastructure. The library is licensed under Apache 2.0.
NannyML is based in Leuven, Belgium. Soda, a data quality company in Brussels, announced its acquisition of NannyML in June 2025 and stated that the open-source project remains maintained. The library runs wherever you install it, so data stays in your environment. According to the documentation, no data leaves your environment in the managed-application option; the public pages do not state a hosting region for the SaaS option. For a European organisation it is an EU-based, self-hostable alternative to US-based ML observability platforms such as Evidently, Arize or Fiddler.