Context. In addition to the discovery of hundreds of exoplanets, the high-precision photometry from the CoRoT and Kepler satellites has led to measurements of surface rotation periods for tens of thousands of stars, which can potentially be used to infer stellar ages via gyrochronology. Aims. Our main goal is to derive ages of thousands of field stars using consistent rotation period measurements derived by different methods. Multiple rotation periods are interpreted as surface differential rotation (DR). We study the dependence of DR with rotation period and effective temperature. Methods. We reanalyze a previously studied sample of 24 124 Kepler stars using different approaches based on the Lomb-Scargle periodogram. Each quarter (Q1-Q14) is treated individually using a prewhitening approach. Additionally, the full time series and their different segments are analyzed. Results. For more than 18 500 stars our results are consistent with the rotation periods from McQuillan et al. (2014, ApJS, 211, 24). Of these, more than 12 300 stars show multiple significant peaks, which we interpret as DR. Dependencies of the DR with rotation period and effective temperature could be confirmed, e.g., the relative DR increases with rotation period. Gyrochronology ages between 100 Myr and 10 Gyr were derived for more than 17 000 stars using different gyrochronology relations, most of them with uncertainties dominated by period variations. We find a bimodal age distribution for T eff between 3200-4700 K. The derived ages reveal an empirical activity-age relation using photometric variability as stellar activity proxy. Additionally, we found 1079 stars with extremely stable (mostly short) periods. Half of these periods may be associated with rotation stabilized by non-eclipsing companions, the other half might be due to pulsations. Conclusions. The derived gyrochronology ages are well constrained since more than ∼93.0% of the stars seem to be younger than the Sun where calibration is most reliable. Explaining the bimodality in the age distribution is challenging, and limits accurate stellar age predictions. The relation between activity and age is interesting, and requires further investigation. The existence of cool stars with almost constant rotation period over more than three years of observation might be explained by synchronization with stellar companions, or a dynamo mechanism keeping the spot configurations extremely stable.