Engineered proteins capable of binding and transporting nucleic acids hold significant potential for advancing disease control in both the medical and agricultural fields. However, identifying small nucleic acid-binding domains remains challenging, as existing predictors primarily classify entire proteins as binders or nonbinders rather than targeting specific binding regions. Here, we introduce NABhClassifier, a highly efficient and precise web server designed to detect small helical sequences with nucleic acid-binding potential. Featuring an intuitive interface and a fully automated prediction pipeline, NABhClassifier integrates eight machine learning models for rapid analysis, delivering results in seconds per protein sequence. Predictions are summarized in the NABh index, a consensus score that combines outputs from all models for enhanced reliability. The server's accuracy has been validated on data sets of DNA-binding and single- and double-stranded RNA-binding proteins from various species. NABhClassifier provides a powerful tool for identifying small helices with nucleic acid-binding capacity, facilitating the discovery of novel biotechnological applications. The server, along with tutorials, is freely accessible at http://143.54.25.149.