Schematic of our atmospheric retrieval framework CROCODILE. The input of the routine is the data – which can be photometry, low-, or medium-resolution spectroscopy –, the forward model, the opacity database, and the priors. CROCODILE starts by sampling parameters from the prior, from which the forward model is evaluated. It is then compared to the data by computing the sum of the log-likelihood functions corresponding to the three observing techniques, from which new parameters are sampled and the loop restarts. The output is the posterior distribution of the forward model.