The COMPASS experiment took data at the M2 beam line at CERN with a rich hadron physics program for the last two decades. This talk focuses on the spectroscopy program, where a 190 GeV negatively charged hadron beam was shot on a fixed proton target.

We investigate diffractive dissociation events in various final states using both the $\pi^-$ and $K^-$ component of the beam. Performing a partial-wave decomposition and fitting the result with resonance models allows us to measure $a_J,\pi_J,K_J$ and $K^*_J$ states and to extract their resonance parameters.

Especially exciting are exotic mesonic states beyond the $q\bar{q}$ systematics of up, down and strange quarks. An example is a hybrid state where gluonic degrees of freedom contribute to the quantum numbers. Phenomenological models as well as recent lattice QCD simulations predict the lightest hybrid state to have spin-exotic quantum numbers $J^{PC}=1^{-+}$. A coupled-channel fit to the COMPASS data using the selected $\eta\pi$ and $\eta^\prime\pi$ final states results in $1^{-+}$ signals at different masses originating from a single pole at about 1.6 GeV/${c}^2$.

In addition, we show the recently released results for the $\omega\pi^-\pi^0$ final state, confirming the $\pi_1(1600)$ state, as well as several other known states.

The strange-light sector is probed with the reaction $K^- p \rightarrow K^-\pi^-\pi^+p$, where COMPASS collected the largest data set to date and we show results for known states as well as new states, e.g. an exotic $0^{-}$ candidate.

Supported by BMBF.