Processing of input data for simulation models can be a major effort. The Python library PyModelData provides many features to aid the model programmer as well as the model user. User input is transferred directly into data structures of the programming language Python without the need to write code. PyModelData is based on YAML which offers an input format readable by humans and machines alike. PyModelData extends this format to allow nesting of files, thus separating frequently changed from other data. Moreover, the included files can have other formats such as CSV, Excel, dBase or HDF.An application user writes an input data file in YAML format while an application programmer may add a declarative input description with units, valid ranges and other meta-information which is applied after parsing the user's data. PyModelData has been used successfully for two simulation models. Users seem to grasp the library quickly and become productive after a short learning period. It is planned to add a GUI interface that generates its views automatically from the declarative template file. This means getting a simple yet useful GUI without any work except specifying the input data.