Starting in December 2014, Kepler K2 observed Neptune continuously for 49 days at a 1-minute cadence. The goals consisted of studying its atmospheric dynamics (Simon et al. 2016), detecting its global acoustic oscillations (Rowe et al., submitted), and those of the Sun, which we report on here. We present the first indirect detection of solar oscillations in intensity measurements. Beyond the remarkable technical performance, it indicates how Kepler would see a star like the Sun. The result from the global asteroseismic approach, which consists of measuring the oscillation frequency at maximum amplitude nu_max and the mean frequency separation between mode overtones Delta nu, is surprising as the nu_max measured from Neptune photometry is larger than the accepted value. Compared to the usual reference nu_max_sun = 3100 muHz, the asteroseismic scaling relations therefore make the solar mass and radius appear larger by 13.8 +/- 5.8 % and 4.3 +/- 1.9 % respectively. The higher nu_max is caused by a combination of the value of nu_max_sun, being larger at the time of observations than the usual reference from SOHO/VIRGO/SPM data (3160 +/- 10 muHz), and the noise level of the K2 time series, being ten times larger than VIRGOs. The peak-bagging method provides more consistent results: despite a low signal-to-noise ratio (SNR), we model ten overtones for degrees l=0,1,2. We compare the K2 data with simultaneous SOHO/VIRGO/SPM photometry and BiSON velocity measurements. The individual frequencies, widths, and amplitudes mostly match those from VIRGO and BiSON within 1 sigma, except for the few peaks with lowest SNR.