Chromatin regulators and associated histone and DNA modifications play a critical role in modulating gene expression states and establishing long-term epigenetic memory. This system is critical in development, aging, and disease, and could provide essential capabilities for synthetic biology. In all these contexts, the temporal dynamics and cell-to-cell variability of gene expression are critical, but have been difficult to study because current methods usually provide static correlations between chromatin modifications and gene expression, and aggregate data across potentially heterogeneous cell populations. Therefore, it has remained unclear how much and how fast does expression change upon recruitment of a specific chromatin regulator, and how permanent are these changes once it unbinds. In order to address these questions, we developed a system that allows us to systematically recruit chromatin regulators and measure their effects on gene expression in real time, at the single-cell level. Using this platform, we analyzed the dynamic effects of four regulators associated with diverse modifications: DNA methylation, histone deacetylation, and histone methylation. We found that gene silencing and subsequent reactivation both occurred through abrupt, all-or-none, and stochastic events. Remarkably, the effects of all regulators could be quantitatively described by a unified model based on stochastic transitions among three discrete states: actively expressing, reversibly silent, and irreversibly silent, with each regulator generating a distinct set of transition rates. These specific transition rates enabled distinct timescales of silencing (hours to days), and types of epigenetic memory (transient, permanent, and a hybrid of the two). As a result, chromatin regulators act as fractional control devices: they take the duration or strength of a signal as input, and use it to control the fraction of cells in which the gene is active, reversibly silenced, or irreversibly silenced. These results provide a predictive, statistical framework for understanding and engineering mammalian chromatin regulation.