Embryonic stem cells co-express Oct4 and Oct1, a related protein with similar DNA-binding specificity. To study the role of Oct1 in ESC pluripotency and transcriptional control, we constructed germline and inducible-conditional Oct1-deficient ESC lines. ESCs lacking Oct1 show normal appearance, self-renewal and growth but manifest defects upon differentiation. They fail to form beating cardiomyocytes, generate neurons poorly, form small, poorly differentiated teratomas, and cannot generate chimeric mice. Upon RA-mediated differentiation, Oct1-deficient cells induce lineage-appropriate developmentally poised genes poorly while lineage-inappropriate genes, including extra-embryonic genes, are aberrantly expressed. In ESCs, Oct1 co-occupies a specific set of targets with Oct4, but does not occupy differentially expressed developmental targets. Instead, Oct1 occupies these targets as cells differentiate and Oct4 declines. These results identify a dynamic interplay between Oct1 and Oct4, in particular during the critical window immediately after loss of pluripotency when cells make the earliest developmental fate decisions. DOI: http://dx.doi.org/10.7554/eLife.20937.001, eLife digest Humans and most other animals are composed of hundreds of different types of cell, including nerve cells, muscle cells and blood cells. Despite performing many different roles, these cells all develop from a single fertilized egg, which divides to make a particular group of cells that when studied in the laboratory are called embryonic stem cells (or ESCs for short). The ability of a cell to become a different cell type is defined as “potency”. ESCs are unique because they can specialize into any type of cell present in the adult organism, and they are therefore called “pluripotent”. However, as the embryo develops, its ESCs gradually lose their potency, and become more and more specialized. The activity of a great number of genes must be regulated during the transition from pluripotent to specialized cells, and some of the mechanisms involved in this transition are still unclear. ESCs are known to need a gene-regulating protein called Oct4 to remain pluripotent and Shen, Kang, Shakya et al. now show that a similar protein named Oct1 is essential for their transition to becoming more specialized. When the gene for Oct1 was deleted from mouse ECSs, they behaved largely like “normal” ESCs, but could not properly mature into certain cell types such as heart and nerve cells. Molecular analyses revealed that Oct4 and Oct1 compete to regulate the activity of many common genes with opposing outcomes: Oct4 keeps ESCs pluripotent while Oct1 leads them to specialize. The Oct4 protein is abundant in ESCs and prevails over Oct1, but as the cells mature, the levels of Oct4 drop, and Oct1 takes over in the regulation of their common target genes. Going forward, a better understanding of how ESCs become specialized will help basic research in the laboratory and allow scientists to tackle new questions about how the human body develops and how our organs work. In the longer-term, these findings might also have applications in the field of regenerative medicine, which aims to repair or replace a person’s cells, tissues or organs to improve their health. DOI: http://dx.doi.org/10.7554/eLife.20937.002