scCircle-seq is a single-cell method for profiling extrachromosomal circular DNAs (eccDNAs) and revealing their diversity and complexity in individual cells. The method was validated in normal and cancer cell lines, showing that most eccDNAs vary between cells and are stochastically inherited during cell division. EccDNAs are preferentially produced from chromatin regions enriched in H3K9me3 and H3K27me3 histone marks and are induced during replication stress. Sequencing of eccDNAs and RNA from the same cell revealed no correlation between eccDNA copy number and gene expression levels, except for oncogenes like MYC. scCircle-seq was applied to prostate and breast cancer specimens, revealing cancer-specific eccDNA landscapes and a higher propensity of eccDNAs to form in amplified genomic regions. scCircle-seq is a scalable tool that can be used to dissect the complexity of eccDNAs across different cell and tissue types, expanding their potential for cancer diagnostics. EccDNAs were first identified in the 1960s as double-minute chromosomes in childhood leukemia samples. Since then, they have been detected in multiple species and implicated in various processes, including human tumorigenesis. Three main approaches are available to study eccDNAs: DNA fluorescence in situ hybridization (FISH), bulk whole genome sequencing (WGS), and Circle-Seq. However, these methods have limitations, such as detecting only abundant eccDNAs or averaging diversity and complexity. scCircle-seq is a single-cell adaptation of Circle-Seq that can be applied to both unfixed and fixed cells or nuclei, including those extracted from tumor biopsies. It was validated by comparing single-cell circular DNA profiles from 49 Colo320DM cells with those from a bulk Circle-Seq experiment. The results showed that scCircle-seq can detect eccDNAs with high accuracy and sensitivity. The genomic landscape of eccDNAs in single cells was investigated, revealing that eccDNAs are highly variable at the single-cell level but may form more defined patterns at the population level. scCircle-seq identified four types of eccDNAs: low frequency low uniformity (LFLU), high frequency high uniformity (HFHU), high frequency low uniformity (HFLU), and low frequency high uniformity (LFHU). HFHU eccDNAs were mainly detected in Colo320DM cells and correspond to large, oncogene-containing ecDNAs previously identified by Circle-Seq. scCircle-seq was also used to profile eccDNAs in patient-derived tumor samples, revealing distinct eccDNA landscapes in different tumor types. The results showed that eccDNAs are highly dynamic under replication stress conditions and during cell division, and that their repertoire can be used to accurately cluster cells of the same origin. scCircle-seq was applied to three patient-derived tumor samples, revealing that eccDNAs are highly variable between cellsscCircle-seq is a single-cell method for profiling extrachromosomal circular DNAs (eccDNAs) and revealing their diversity and complexity in individual cells. The method was validated in normal and cancer cell lines, showing that most eccDNAs vary between cells and are stochastically inherited during cell division. EccDNAs are preferentially produced from chromatin regions enriched in H3K9me3 and H3K27me3 histone marks and are induced during replication stress. Sequencing of eccDNAs and RNA from the same cell revealed no correlation between eccDNA copy number and gene expression levels, except for oncogenes like MYC. scCircle-seq was applied to prostate and breast cancer specimens, revealing cancer-specific eccDNA landscapes and a higher propensity of eccDNAs to form in amplified genomic regions. scCircle-seq is a scalable tool that can be used to dissect the complexity of eccDNAs across different cell and tissue types, expanding their potential for cancer diagnostics. EccDNAs were first identified in the 1960s as double-minute chromosomes in childhood leukemia samples. Since then, they have been detected in multiple species and implicated in various processes, including human tumorigenesis. Three main approaches are available to study eccDNAs: DNA fluorescence in situ hybridization (FISH), bulk whole genome sequencing (WGS), and Circle-Seq. However, these methods have limitations, such as detecting only abundant eccDNAs or averaging diversity and complexity. scCircle-seq is a single-cell adaptation of Circle-Seq that can be applied to both unfixed and fixed cells or nuclei, including those extracted from tumor biopsies. It was validated by comparing single-cell circular DNA profiles from 49 Colo320DM cells with those from a bulk Circle-Seq experiment. The results showed that scCircle-seq can detect eccDNAs with high accuracy and sensitivity. The genomic landscape of eccDNAs in single cells was investigated, revealing that eccDNAs are highly variable at the single-cell level but may form more defined patterns at the population level. scCircle-seq identified four types of eccDNAs: low frequency low uniformity (LFLU), high frequency high uniformity (HFHU), high frequency low uniformity (HFLU), and low frequency high uniformity (LFHU). HFHU eccDNAs were mainly detected in Colo320DM cells and correspond to large, oncogene-containing ecDNAs previously identified by Circle-Seq. scCircle-seq was also used to profile eccDNAs in patient-derived tumor samples, revealing distinct eccDNA landscapes in different tumor types. The results showed that eccDNAs are highly dynamic under replication stress conditions and during cell division, and that their repertoire can be used to accurately cluster cells of the same origin. scCircle-seq was applied to three patient-derived tumor samples, revealing that eccDNAs are highly variable between cells