The exhaustive exploration of human cell heterogeneity requires the unbiased identification of molecular signatures that can serve as unique cell identity cards for every cell in the body. However, the stochasticity associated with high-throughput single-cell RNA sequencing has made it necessary to use clustering-based computational approaches in which the transcriptional characterization of cell-type heterogeneity is performed at cell-subpopulation level rather than at full single-cell resolution. We present here Cell-ID, a clustering-free multivariate statistical method for the robust extraction of per-cell gene signatures from single-cell RNA-seq. Cell-ID allows unbiased cell identity recognition across different donors, tissues-of-origin, model organisms and single-cell omics protocols. Cell-ID is distributed as an open-source R software package: https://github.com/RausellLab/CelliD.