RNA genes are ubiquitous in the cell and are involved in a number of biochemical processes. Because there is a close relationship between function and structure, software tools that predict the secondary structure of noncoding RNAs from the base sequence are very helpful. In this article, we focus our attention on the inference of conserved secondary structure for a group of homologous RNA sequences. We present the caRNAc software, which enables the analysis of families of homologous sequences without prior alignment. The method relies both on comparative analysis and thermodynamic information.