Submillimeter interferometry has the potential to image supermassive black holes on event horizon scales, providing tests of the theory of general relativity and increasing our understanding of black hole accretion processes. The Event Horizon Telescope (EHT) performs these observations from the ground, and its main imaging targets are Sagittarius A* in the Galactic Center and the black hole at the center of the M87 galaxy. However, the EHT is fundamentally limited in its performance by atmospheric effects and sparse terrestrial $(u,v)$-coverage (Fourier sampling of the image). The scientific interest in quantitative studies of the horizon size and shape of these black holes has motivated studies into using space interferometry which is free of these limitations. Angular resolution considerations and interstellar scattering effects push the desired observing frequency to bands above 500 GHz. This paper presents the requirements for meeting these science goals, describes the concept of interferometry from Polar or Equatorial Medium Earth Orbits (PECMEO) which we dub the Event Horizon Imager (EHI), and utilizes suitable space technology heritage. In this concept, two or three satellites orbit at slightly different orbital radii, resulting in a dense and uniform spiral-shaped $(u,v)$-coverage over time. The local oscillator signals are shared via an inter-satellite link, and the data streams are correlated on-board before final processing on the ground. Inter-satellite metrology and satellite positioning are extensively employed to facilitate the knowledge of the instrument position vector, and its time derivative. The European space heritage usable for both the front ends and the antenna technology of such an instrument is investigated. Current and future sensors for the required inter-satellite metrology are listed. Intended performance estimates and simulation results are given.