Aims. We analyze the clustering of galaxy clusters in a large contiguous sample, the Constrain Dark Energy with X-ray (CODEX) sample. We construct a likelihood for cosmological parameters by comparing the measured clustering signal and a theoretical prediction, and use this to obtain parameter constraints. Methods. We measured the three multipole moments (monopole, quadrupole, and hexadecapole, ℓ = 0, 2, 4) of the power spectrum of a subset of the CODEX clusters. To fully model cluster clustering, we also determined the expected clustering bias of the sample using estimates for the cluster masses and a mass-to-bias model calibrated using N-body simulations. We estimated the covariance matrix of the measured power spectrum multipoles using a set of simulated dark-matter halo catalogs. Combining all these ingredients, we performed a Markov chain Monte Carlo sampling of cosmological parameters Ωm and σ8 to obtain their posterior. Results. We found the CODEX clustering signal to be consistent with an earlier X-ray selected cluster sample, the REFLEX II sample. We also found that the measured power spectrum multipoles are compatible with the predicted, bias-scaled linear matter power spectrum when the cosmological parameters determined by the Planck satellite are assumed. Furthermore, we found the marginalized parameter constraints of Ωm = 0.24‑0.04+0.06 and σ8 = 1.13‑0.24+0.43. The full 2D posterior is consistent, for example, with the Planck cosmology within the 68% confidence region.