We report on the discovery of a bound exoplanetary microlensing event from a blind search of data gathered from Campaign 9 of the Kepler K2 mission (K2C9). K2-2016-BLG-0005Lb is a densely sampled, binary caustic-crossing microlensing event with caustic entry and exit points that are resolved in the K2C9 data, enabling the lens-source relative proper motion to be measured. We have fitted a binary microlens model to the K2 dataset, and to simultaneous observations from the Optical Gravitational Lensing Experiment (OGLE-IV), Canada-France-Hawaii Telescope (CFHT), Microlensing Observations in Astrophysics (MOA-2), the Korean Microlensing Telescope Network (KMTNet), and the United Kingdom InfraRed Telescope (UKIRT). Whilst the groundbased data only sparsely sample the binary caustic, they provide a clear detection of parallax that allows us to break completely the microlensing mass-position-velocity degeneracy and measure the planet's mass directly. We find a host mass of 0.58 ± 0.03 M and a planetary mass of 1.1 ± 0.1 M J . The system lies at a distance of 5.2 ± 0.2 kpc from Earth towards the Galactic bulge. The projected physical separation of the planet from its host is found to be 4.2 ± 0.3 au which, for circular orbits, corresponds to 𝑎 = 4.4 +1.9 −0.4 au and period 𝑃 = 13 +9 −2 yr, making K2-2016-BLG-0005Lb a close Jupiter analogue. Though previous exoplanet microlensing events have included space-based data, this event is the first bound microlensing exoplanet to be discovered from space-based data. Even through a space telescope not designed for microlensing studies, this result highlights the advantages for exoplanet microlensing discovery that come from continuous, high-cadence temporal sampling that is possible from space. The precision mass and distance measurements of cool low-mass exoplanets from microlensing that will be obtainable from the NASA Nancy Grace Roman Space Telescope and, potentially, the ESA Euclid space telescope, will unlock a key regime for testing planet formation theories.