1. In this paper, through an extension of the N-mixture family of models, we achieve a significant improvement of the statistical properties of the rare species abundance estimators when sample sizes are low, yet of typical size in neotropical bird studies. The proposed method harnesses information from other species in the targeted ecological community to correct each species' estimator. We provide guidance to determine the sample size required to estimate accurately the abundance of rare neotropical bird species. 2. We evaluate the proposed methods using an assumption of 50m fixed radius point count and perform simulations comprising a broad range of sample sizes, true abundances and detectability values. The extension of the N-mixture model is achieved by assuming that the detection probabilities of a set of species are all drawn at random from a beta distribution. This hierarchical model avoids having to specify one detection probability parameter per species in the targeted community. Parameter estimation is done via Maximum Likelihood using data cloning. 3. We compared our methodology with standard N-mixture models, which we show here are severely biased and highly variable when the true abundances of species in the community are less than seven individuals per 100ha. For more common species, the number of point counts and replicates needed to reduce the bias of N-mixture model estimators estimation is high. The beta N-mixture model proposed here outperforms the traditional N-mixture model thus allowing the estimation of organisms at lower densities and control of the bias in the estimation. 4. We illustrate how our methodology can be used to determine the sample size required to estimate the abundance of organisms. We also give practical advice for researchers seeking to propose reliable sampling designs for single species' studies. When the interest is full communities, our model and estimation methodology can be seen as a practical solution to estimate organism densities from rapid inventories datasets. The statistical inferences with this model can also inform differences in ecology and behavior of species when they violate the assumption of a single distribution of detectabilities.