The article introduces a method called Safe-Sequencing System (Safe-SeqS) to improve the detection and quantification of rare mutations using massively parallel sequencing. The key components of Safe-SeqS are assigning unique identifiers (UIDs) to each DNA template molecule, amplifying these molecules to create UID families, and redundant sequencing of the amplification products. A UID family is considered a "supermutant" if at least 95% of its members contain the same mutation. This approach significantly reduces the error rate associated with conventional sequencing methods, making it suitable for detecting rare mutations in various biological samples, including normal human cells and mitochondrial DNA. The authors demonstrate the effectiveness of Safe-SeqS by applying it to measure the fidelity of DNA polymerases, analyze oligonucleotide compositions, and determine the prevalence of mutations in the *CTNNB1* gene and mitochondrial DNA. The results show that Safe-SeqS can reduce the apparent mutation frequency by several orders of magnitude, providing a more accurate and reliable method for detecting rare mutations.The article introduces a method called Safe-Sequencing System (Safe-SeqS) to improve the detection and quantification of rare mutations using massively parallel sequencing. The key components of Safe-SeqS are assigning unique identifiers (UIDs) to each DNA template molecule, amplifying these molecules to create UID families, and redundant sequencing of the amplification products. A UID family is considered a "supermutant" if at least 95% of its members contain the same mutation. This approach significantly reduces the error rate associated with conventional sequencing methods, making it suitable for detecting rare mutations in various biological samples, including normal human cells and mitochondrial DNA. The authors demonstrate the effectiveness of Safe-SeqS by applying it to measure the fidelity of DNA polymerases, analyze oligonucleotide compositions, and determine the prevalence of mutations in the *CTNNB1* gene and mitochondrial DNA. The results show that Safe-SeqS can reduce the apparent mutation frequency by several orders of magnitude, providing a more accurate and reliable method for detecting rare mutations.