Runtime monitoring is an essential part of guaranteeing the safety of cyber-physical systems. Recently, runtime monitoring frame- works based on formal specification languages gained momentum. These languages provide valuable abstractions for specifying the behavior of a system. Yet, writing specifications remains challenging as, among other things, the specifier has to keep track of the timing behavior of streams. The stream-based specification language RTLola has a rich type system and several other static analysis steps that provide additional insights about the specification. So far, syntactic and semantic errors are only available when locally executing the RTLola front end. This means that all the information provided by the front end is not readily available while editing and can therefore not be used to better understand the specification or to spot errors. Thus, this thesis describes the new web-based IDE RTLola Playground and how it integrates and visualizes the information provided by the RTLola front end. Different approaches of handling specifications with syntactic or semantic errors are also discussed. The RTLola Playground is evaluated in a small user study, and proves that it helps with detecting intention errors, meaning a mismatch between the specification and the intended behavior, and understanding in general. In addition, a different design of the front end is explored with the aim to be more incremental and being able to provide more information about syntactically or semantically invalid specifications.