Viruses are a critical component of biosphere and human ecosystems, directly influencing human health and disease and controlling the output of engineered ecosystems. Unfortunately for scientists studying them, most viral signals remain hidden in sequence datasets due to a lack of optimal identification tools. A new tool aims to make virus identification easier. VirSorter2 is a DNA and RNA virus identification tool that leverages genome-informed database advances to improve the accuracy and range of virus sequence detection. Compared with other tools, VirSorter2 performs more consistently with high accuracy across diverse viruses and minimized errors associated with atypical cellular sequences, including eukaryotic genomes and plasmids. VirSorter2 is also modular, allowing it to expand to newly discovered viruses through the design of new viral classifiers. Overall, VirSorter2 demonstrates higher accuracy across major viral groups, advancing our knowledge of virus evolution, diversity and virus-microbe interaction in various ecosystems. VirSorter2 is available both on bioconda and as an iVirus app at https://de.cyverse.org/de.