The lowest metallicity stars are also the oldest ones and they carry the imprint of the first supernovae. Since the very first stars are likely short-lived and inaccessible to us, the lowest metallicity stars are also those that can inform us most on the first generation of stars, how they enriched their environment, and produced the first structures that, through hierarchical formation, built up the primordial constituents of the galaxies we observe today. These stars are exceedingly rare and few of them are known today. In order to efficiently improve on this situation, we have put in place the Pristine international collaboration. Focussed around a wide narrow-band photometric survey conducted at the Canada-France-Hawaii Telescope and a large dedicated spectroscopic campaign, Pristine is many times more efficient than previous attempts at finding the precious low-metallicity stars. From the detailed study of these stars, we will: hunt for the most extreme low-metallicity stars; place constraints on star formation during the earliest epoch of the universe; reliably unveil the properties of the faintest known dwarf galaxies that orbit the Milky Way and are promising cosmological probes; decompose the Milky Way into its main components to place our Galaxy in the global context of galaxy formation and evolution; deconstruct the stellar halo of the Milky Way into its constituent substructures, which will then be used to constrain the mass and shape of the Milky's Way potential. In order to reliably achieve these significant goals, we request funding to support Pristine in France and ensure the project is staffed adequately to yield high-impact scientific returns in the field of very low metallicity stars. Our team is built around experts of the field in Nice (OCA/Lagrange), Paris (GEPI), and Strasbourg (Observatoire astronomique de Strasbourg) and represents a large part of the full Pristine collaboration. To complement this team, we ask for funding for a PhD student and two postdoctoral researchers who will be spread over the three French nodes of our project, along with funds to support the team and secure the visibility and active partnership of the French team members within the full international Pristine collaboration. We wish to emphasize that the effort already invested by the current team members into preparing the survey (successful telescope time proposals, data acquisition, reduction, and calibration, start of the spectroscopic follow-up campaign) means that the Pristine project is a low risk but high return project if it were to be supported by the ANR. It builds on large facilities and surveys with a significant French involvement (CFHT, Gaia, WEAVE) and it promises numerous high impact papers to be published in the high-visibility fields of Galactic archaeology and near-field cosmology in which France is a world leader.