Gaia is observing the position and path of an enormous number of stars. Future experiments promise to do so for even more stars at high precision. If the light from such a star is traveling to us through a stochastic gravitational wave background, the apparent position of the star will be displaced by an amount essentially independent of the distance but dependent on the strength of the gravitational wave. Correlations in such shifts from many stars can be used to detect or constrain a gravitational wave background. The frequency range probed partially fills the gap between pulsar timing arrays and LISA. We explore the potential cosmological consequences of this by looking at phase transitions, topological defects and pbhs.