Zingg is an open-source tool for entity resolution, identity resolution and master data management (MDM). Using machine learning, it learns which records refer to the same person or organisation, so duplicate customer or supplier profiles in a data lake or warehouse can be identified. You train the model by labelling a few dozen record pairs as matches or non-matches with an interactive tool. Zingg assigns a persistent identifier (ZINGG_ID) and can update clusters incrementally. It runs on Apache Spark, works with platforms such as Databricks, Snowflake, BigQuery, Redshift and Microsoft Fabric, and can be started from a Docker image.
The project is released under the AGPL-3.0 licence as a free Community edition on GitHub; a paid Enterprise edition also exists. According to its website, Zingg Labs, Inc. is based in Lewes, Delaware, in the United States, so it is a US vendor. Because Zingg runs on any private or cloud-based Spark service, you can operate it in your own environment, including in the EU, so that data stays where your own platform keeps it.