Current and future large surveys will produce unprecedented amounts of data. Realistic simulations have become essential for the design and development of these surveys, as well as for the interpretation of the results. We present MAMBO, a flexible and efficient workflow to build empirical galaxy and active galactic nucleus (AGN) mock catalogues that reproduce the physical and observational properties of these sources. We started with simulated dark matter (DM) haloes, to preserve the link with the cosmic web, and we populated them with galaxies and AGN using abundance matching techniques. We followed an empirical methodology, using stellar mass functions, host galaxy AGN mass functions, and AGN accretion rate distribution functions studied at different redshifts to assign, among other properties, stellar masses, the fraction of quenched galaxies, or the AGN activity (demography, obscuration, multiwavelength emission, etc.). As a proof test, we applied the method to a Millennium DM lightcone of 3.14 $ deg^2$ up to a redshift of $z=10$ and down to stellar masses of $ M \, M_ We show that the AGN population from the mock lightcone presented here reproduces with good accuracy various observables, such as state-of-the-art luminosity functions in the X-ray up to $z 7$ and in the ultraviolet up to $z 5$, optical/near-infrared colour-colour diagrams, and narrow emission line diagnostic diagrams. Finally, we demonstrate how this catalogue can be used to make useful predictions for large surveys. Using Euclid as a case example, we compute, among other forecasts, the expected surface densities of galaxies and AGN detectable in the Euclid $H_ E $ band. We find that Euclid might observe (on $H_ E $ only) about $10^ $ and $8 $ type 1 and 2 AGN, respectively, and $2 $ galaxies at the end of its $ deg^2$ Wide survey, in good agreement with other published forecasts.