Requirements for software applications change almost continuously in all phases of the product life cycle, which leads to a constant adaption of software systems. These changes affect the domain model of the application and usually lead to necessary changes of the underlying database. As a result, the data stored in this database must also be adapted accordingly. This well-known problem is called schema evolution. Manual schema evolution with low-level SQL scripts contains both time-consuming and error-prone routine tasks and complicated operations that require expert knowledge. Automation promises significant leverage for cost savings and quality improvements. To date, the most advanced solutions for this problem have focused on operator-based approaches. However, these approaches do not allow free editing of the database schema, but require the database administrator to know in advance the exact sequence of all required operations. In this paper, we present EvolveDB, an approach to model-driven schema evolution in relational databases, where the user specifies the evolution steps by freely editing a database model extracted by reverse engineering. EvolveDB analyzes the differences between the status quo and the evolved model structures and generates a data migration script. An initial evaluation with experts shows promising results. A screencast of the demo is available at https://youtu.be/ieXmrDd2nw4.