We introduce the SAPP (Stellar Abundances and atmospheric Parameters Pipeline), the prototype of the code that will be used to determine parameters of stars observed within the core program of the PLATO space mission. The pipeline is based on the Bayesian inference and provides effective temperature, surface gravity, metallicity, chemical abundances, and luminosity. The code in its more general version has a much wider range of potential applications. It can also provide masses, ages, and radii of stars and can be used with stellar types not targeted by the PLATO core program, such as red giants. We validate the code on a set of 27 benchmark stars that includes 19 FGK-type dwarfs, 6 GK-type subgiants, and 2 red giants. Our results suggest that combining various observables is the optimal approach, as this allows the degeneracies between different parameters to be broken and yields more accurate values of stellar parameters and more realistic uncertainties. For the PLATO core sample, we obtain a typical uncertainty of 27 (syst.) +/- 37 (stat.) K for T-eff, 0.00 +/- 0.01 dex for log g, 0.02 +/- 0.02 dex for metallicity [Fe/H], -0.01 +/- 0.03 R-circle dot for radii, -0.01 +/- 0.05 M-circle dot for stellar masses, and -0.14 +/- 0.63 Gyr for ages. We also show that the best results are obtained by combining the nu(max) scaling relation with stellar spectra. This resolves the notorious problem of degeneracies, which is particularly important for F-type stars.