The derivative-enabled simulation plugin API, FSMDAO, is presented for the FlowSimulator HPC ecosystem in order to enable gradient-based MDAO with state-of-the-art Python frameworks like OpenMDAO or GEMSEO. This allows FlowSimulator plugins to be orchestrated by such a framework. By means of FSMDAO, plugins systematically expose their inputs, outputs and partial derivatives thereof, including derivatives obtained by forward/reverse algorithmic differentiation (AD). Implicit plugin components additionally offer the residuals of the discretized nonlinear and linearized governing equations together with block inversion capabilities for both nonlinear and linearized problems. Large mesh datasets like input and output vectors required for the multidisciplinary coupling are shared among the simulation plugins in memory via the FlowSimulator Data Manager (FSDM) in an MPI-parallel way. The individual FlowSimulator plugins often decompose into a number of implicit/explicit FSMDAO plugins separating the core solver part and post processing components so that MDAO workflows can be set up and executed in a modular and granular manner. Additionally, a standardized API for simulation plugins allows for standardized testing of the simulation software of FlowSimulator ecosystem, e.g. verifying the derivatives of plugin against finite difference approximations. A selection of coupled workflows will be discussed in the presentation to explain the MDAO framework integration via FSMDAO. Among the simulation plugins involved are the CFD Software by ONERA, DLR and Airbus (CODA), the general-purpose FE structure mechanics software b2000++, the open-source CAD kernel OCCT or a virtual-blade body-force propeller model coupled to the CFD simulation. Based on the examples, the integrated framework approach will be evaluated and assessed in conjunction with MDAO algorithms provided by OpenMDAO in terms of plugin integration, usability and scalability properties. An outlook will be provided regarding the potential and challenges for large-scale, gradient-based MDAO using the suggested software stack.