Distances to stars are essential for studying our Galaxy's formation history. However, they are difficult to measure precisely, especially for stars at large distances. Therefore, pristine samples of red clump stars are highly sought after given that they are standard candles and give precise distances even at large distances. However, it is difficult to cleanly select red clump stars because they can have the same atmospheric parameters (e.g., \teff\ and \logg) as red giant branch stars which are not standard candles. Recently, it was shown that the asteroseismic parameters, \ps\ and \fsep, which are used to accurately select red clump stars, can be derived from spectra using the change in the surface carbon to nitrogen ratio ([C/N]) caused by mixing during the red giant branch phase. In this study, we show this change in [C/N] can also impact the spectral energy distribution. This enables us to predict the values of \ps, \fsep, \teff\ and \logg\ using 2MASS, AllWISE, \gaia, and Pan-STARRS data in order to select a clean sample of red clump stars for the first time. We achieve a contamination rate of $\sim$20%, equivalent to that achieved when selecting based on \teff\ and \logg\ values derived from low resolution spectra. Finally, we present two red clump samples. One sample has a contamination rate of ∼ 20% and ∼ 405,000 red clump stars. The other has a contamination of ∼ 33% and ∼ 2.6 million red clump stars which includes ∼ 75,000 stars at distances > 10 kpc. For |b$|>$30 degrees we find ∼ 15,000 stars with a contamination rate of ∼ 9%. The scientific potential of this catalog for studying the structure and formation history of the Galaxy is vast given that it provides millions of precise distances to stars in the inner bulge and distant halo where astrometric distances are imprecise.