This talk presents our current project to deliver performance modeling and prediction capabilities as a service. In order to do so we have developed a scalable web application architecture that allows to deliver these capabilities to as many clients as needed by allowing to scale out easily. To implement these capabilities, we have leveraged frameworks such as Polymer for the user interface, REST services for the backend, Cassandra as data storage and novel cloud features such as Amazon Lambda for distributing simulation jobs. The underlying performance modeling notation for the solution is derived from the Palladio Component Model (PCM). The talk will also feature a demo of the solution. In the demo, we will explain why we have hidden a lot of the features of PCM from the user to focus on the most important use cases and thereby improving the user experience.