Context: With certain assumptions, radiative energy transport can be modelled by the diffusion approximation. In this case, the Rosseland mean opacity coefficient characterises the interaction between radiation and matter. The opacity data are usually available in pre-tabulated form, and in the generation of the data one assumes a distinct heavy element mixture, which is usually a scaled solar one. Therefore, presently available data is unable to cover the full parameter range of some astrophysical problems, in which the chemical composition of the medium being considered varies. Aims: We attempt to produce low temperature opacity data incorporating the effects of varied abundances of the elements carbon and nitrogen. For our temperature range of interest, molecules represent the dominant opacity source. Our dataset covers a wide metallicity range and is meant to provide important input data for stellar evolution models and other applications. Methods: We conduct chemical equilibrium calculations to evaluate the partial pressures of neutral atoms, ions, and molecules. Based on a large dataset containing atomic line and continuum data and, most importantly, a plethora of molecular lines, we calculate Rosseland mean opacity coefficients not only for a number of different metallicities, but also for varied abundances of the isotopes 12C and 14N at each metallicity. The molecular data comprise the main opacity sources for either an oxygen-rich or carbon-rich chemistry. We tabulate the opacity coefficients as a function of temperature and, basically, density. Results: Due to the special role of the CO molecule, within a certain chemistry regime an alteration to the carbon abundance causes considerable changes in the Rosseland opacity. The transition from a scaled solar (i.e. oxygen-rich) mixture to a carbon-rich regime results in opacities that can, at low temperatures, differ by orders of magnitude from to the initial situation. The reason is that the mean opacity in either case is due to different molecular absorbers. Variations in the abundance of nitrogen have less pronounced effects but, nevertheless, cannot be neglected. Conclusions: In typical astrophysical applications, it is indispensable to take into account opacity variations due to chemistry changes. In this respect, the new data is superior to previous compilations, but is, however, still subject to uncertainties.