Currently, cloud optical depth is one of the most poorly observed climate variables. The new “cloud mode” in the Aerosol Robotic Network (AERONET) will inexpensively yet dramatically increase cloud optical depth observations in both number and accuracy. We tested cloud mode retrievals at the Atmospheric Radiation Measurement program’s Oklahoma site for a variety of situations ranging from broken clouds to overcast. For overcast cases, 1.5-min average cloud-mode cloud optical depth retrievals agreed to better than 15% with those from a standard ground-based flux method. For broken clouds, we found that our retrievals could capture the rapid variations revealed by the microwave radiometer. A three-year climatology of (daytime) cloud-mode retrievals agreed reasonably well with cloud radar retrievals both in seasonal variation and occurrence distribution of optical depth. However, our monthly-mean cloud optical depths are generally larger than cloud radar retrievals, mainly due to the current observation strategy of cloud mode that excludes optically thin clouds. Cloud products from AERONET not only demonstrate a unique approach to acquire cloud and aerosol properties using a single instrument, but also greatly enhance current cloud observations and expand them to a global scale.