New release of synthetic data platform incorporates customer feedback, role-based access and MySQL support 

Vienna, Austria/New York, NY – May, 19, 2022 - MOSTLY AI today launched the latest generation of its industry-leading synthetic data generator. In response to customer feedback, MOSTLY AI 2.2 now includes a variety of features and capabilities that deliver more speed, agility and resilience while using less hardware resources and providing even further improved out-of-the-box privacy security.

Role-based access

As the number of people using synthetic data within organizations grows, the need for different access levels becomes essential. A data science team for instance will need a different environment and configurations than a software testing team.Thanks to the new release, administrators can now set up different teams and user roles to serve wildly different use cases within the same company. Sharing synthetic data between various teams is also possible with MOSTLY AI 2.2, making life easier for everyone.

Scale and velocity

MOSTLY AI’s synthetic data generator can not only process datasets virtually limitless in size. The brand-new MOSTLY AI 2.2 is now faster than ever before: The time to train the AI model has halved, and there is a ten-fold increase in synthetic data generation performance. What the platform used to generate in minutes can now be done in seconds. 

MySQL data connector

MOSTLY AI 2.2 now supports one of the most popular database management systems used in cloud environments. The MySQL data connector enables synthetic data in the cloud and integrates MOSTLY AI with cloud databases like AWS Aurora, Google Cloud SQL, and many more. Apart from the newly released MySQL integration, the platform also supports MS SQL, Oracle, PostgreSQL, and Db2. 

Synthetic data with privacy by design

The synthetic data generation has become even smarter and safer with the 2.2 release. MOSTLY AI protects rare categories by replacing them with non-rare categories in a context-aware manner. Further improved privacy measures, such as extreme sequence length protection and extreme value protection in all numerical formats, make it impossible to create a synthetic dataset that leaks privacy.