acados/acados ? reverse-engineered prompt

Build me a fast C library for nonlinear optimal control and model predictive control that can run in real time on embedded systems.

I want it to solve optimal control problems and moving horizon estimation problems efficiently, with support for nonlinear dynamics, economic cost functions, algebraic constraints, multiple shooting, and both ODE and DAE models. It should be modular, so different solver and integration choices can be combined, and it should expose easy to use interfaces for Python, MATLAB, and Octave, ideally through a CasADi style workflow for defining models and constraints.

Please include a clean build setup, examples, tests, and documentation so someone can install it, run a demo, and understand how to use it for MPC and estimation. If you need current details from the docs or related projects, look them up online and follow the existing project style.