We use machine learning techniques on textual data to identify financial crises. The onset ofa crisis and its duration have implications for real economic activity, and as such can be valuableinputs into macroprudential, monetary, and fiscal policy. The academic literature and the policyrealm rely mostly on expert judgment to determine crises, often with a lag. Consequently, crisisdurations and the buildup phases of vulnerabilities are usually determined only with the benefitof hindsight. Although we can identify and forecast a portion of crises worldwide to variousdegrees with traditional econometric techniques and using readily available market data, we findthat textual data helps in reducing false positives and false negatives in out-of-sample testingof such models, especially when the crises are considered more severe. Building a frameworkthat is consistent across countries and in real time can benefit policymakers around the world, especially when international coordination is required across different government policies.