I jointly estimate monthly series for GDP and eight subcomponents for the US since 1950. The series match 1) quarterly national accounts equivalents, 2) exact data on monthly consumption, and 3) past relationships with other monthly indicators. I estimate the Kalman filter parameters by GMM, allowing fast calculation of confidence intervals for monthly estimates including parameter uncertainty, and validate the confidence intervals. After 1970 standard errors are tight, less than 0.3pp of GDP, and point estimates informative, with standard deviations four times the standard error. I provide confidence intervals for recessions and show that output peaks line up well with the onset of NBER recessions, but troughs often predate NBER equivalents.