Context. Diagrams are an important artifact in software engineering. They are used to visualize complex systems and for communication with different stakeholders, including domain experts and developers. Problem. Currently, two major diagramming approaches exist: either the diagram is created using a textual concrete syntax, or using a graphical editor. However, both have their limitations, in particular when precise manual layouting is required. In the area of modeling, recent research introduces the concept of hybrid/blended modeling, which combines graphical and textual notations for an improved user experience. Our goal is to apply this concept to diagramming, too. Objective. Our concept allows manipulating the layout both graphically and textually, with both views being live-synced. Also, for improved clarity of complex layouts, we support programming language constructs in the textual definition. First, we want to collect and evaluate the requirements of a hybrid textual-graphical diagramming approach, with a focus on the interactive graphical view. Second, we want to prove that a concept of modular hybrid diagramming is feasible. Method. In order to collect requirements, we perform interviews with experts. Then, we perform a survey with the same experts to evaluate the collected requirements. To prove our concept is feasible, we implement a modular framework for hybrid diagramming. To evaluate it, we implement a module for UML class diagrams. Result. Overall, we succeed in implementing our modular hybrid diagramming framework and thus show that our approach is feasible. By conducting two case studies, we show that our implementation results in a usable diagramming experience for UML class diagrams. Conclusion. We developed a tool that allows for hybrid textual-graphical creation of UML class diagrams using a web-based editor. However, future work is required to implement missing features, especially in the area of graphical editing.