Nowcasts are tools used to provide timely and accurate water-quality assessments of threats to drinking-water and recreational resources from fecal contamination or cyanobacterial harmful algal blooms. They use mathematical models and techniques to provide near-real-time estimates of fecal-indicator bacteria (FIB) and cyanotoxin concentrations. Techniques include logic-based thresholds, decision trees (built with machine learning), multiple linear and binary logistic regression, artificial neural networks, and process-based deterministic models. The type of site (freshwater, marine, or river) and dependent variable (FIB or cyanotoxin) dictate which explanatory variables are used in models. Nowcast systems notify the public of associated public-health risks and can also be used to manage data for FIB models; work is ongoing to incorporate cyanotoxin models into some nowcasts. The Great Lakes NowCast in the USA has been operational since 2010 and includes 25 lake beaches and one recreational river site. Examples of other operational FIB nowcasts are described for locations in the USA and around the world. In many cases, models predicted exceedances of FIB standards with accuracies as good as or better than using the previous measured FIB concentration (persistence method). Accuracy and timeliness are vital to beach management decisions that protect public health and support the local recreation-driven economy. Nowcasts benefit the public by providing estimates of water-quality conditions in near-real-time. Managers can use nowcasts at recreational and drinking-water treatment plant sites when FIB or cyanotoxins are projected to be elevated to target sample collection, to provide near-real-time recreational advisories to the public, or to preemptively optimize drinking-water treatments or change intake options to mitigate possible adverse effects on drinking-water quality.