The art of X-ray production from relativistic electron beams has undergone several remarkable advances during the past several decades, revolutionizing structural and dynamical studies of atomic and molecular aggregates for science and technology for the 21st century. The storage ring-based sources advanced from the 2GSR (2nd Generation Storage Ring) to the 3GSR with optimized ring design and the use of undulators. Currently, we are in the doorstep into the 4GSR with a hundredfold higher brightness and the “diffraction limited” hard X-rays. In an X-ray free-electron laser (XFEL), the radiation amplitudes from individual electrons add coherently leading to a large enhancement of intensities. The LCLS at SLAC was the first XFEL that became operational a decade ago. Similar XFELs have been built since then at several accelerator laboratories around the world, including the PAL XFEL in Korea. The latest is the European XFEL, employing a large pulsed super-conducting linac. The XFELs so far have been high-gain, single-pass devices, amplifying the initial noise to intense, quasi-coherent X-ray pulses known as self-amplified spontaneous emission (SASE). The SASE pulses can be femtoseconds or shorter, suitable for ultrafast dynamics or imaging of large molecules via “diffraction before destruction”. A low-gain XFEL is also possible by employing an X-ray cavity formed by Bragg reflectors in which a stored X-ray pulse is repeatedly amplified. This so-called X-ray FEL Oscillator (XFELO) is currently under development, promising to become a “real X-ray laser.” In addition to being fully coherent, it will also provide an ultra-narrow bandwidth of a few milli-eV. With a further stabilization, an XFELO can produce X-ray spectral comb for fundamental sciences.