Google-Microsoft-OSM Open Buildings (VIDA)
Google-Microsoft-OSM Open Buildings is a single global layer of building footprints that VIDA assembled by conflating three existing datasets: Google Open Buildings V3, Microsoft Global ML Footprints, and OpenStreetMap. Together they total 2,705,459,584 footprints, and every footprint is tagged with the provider it came from, so you can tell a satellite-derived ML polygon from a community-mapped OSM building. Because it stitches the three sources into one deduplicated layer, coverage is more complete than any single provider reaches alone, which matters most in regions where one dataset is thin.
It is hosted on Source Cooperative (a Radiant Earth project) and distributed in cloud-native formats: GeoParquet, FlatGeobuf, and PMTiles. The data is partitioned by country and by S2 cell, so you can pull one nation or a small tile rather than the whole planet, and you can query the GeoParquet directly with DuckDB or load the PMTiles straight into a web map. A global PMTiles layer is provided for visualisation.
For RAG and AI this is a geometry and location asset rather than a text corpus. Join footprints to addresses, place names, or population figures, turn each building or neighbourhood into a short descriptive record, and embed those records so a model can answer questions grounded in the built environment: building density, coverage gaps, or which structures sit inside an area. It suits mapping, urban-planning, logistics, and disaster-response work.
On licensing, the combined dataset is released under the Open Data Commons Open Database Licence (ODbL) v1.0. That permits commercial use but is share-alike: if you publicly distribute a database derived from it you must license that database under the same terms and credit the sources, so keep the provider labels. Compared with our OpenStreetMap and Overture Maps entries this one is narrower (buildings only) but broader in footprint coverage, and it is the better starting point when completeness of building geometry is the priority.
Related sources
GeoNames
A geographical database of more than 11 million place names, each with coordinates, population, elevation, and administrative divisions, covering every country. It works well as a gazetteer for resolving place names to locations.
Natural Earth
A public domain map dataset built for cartography, offered at three scales: 1:10m, 1:50m, and 1:110m. It bundles cultural data (borders, cities, roads), physical data (coastlines, rivers, lakes), and raster imagery, all ready to drop into a map.
OpenStreetMap
A crowdsourced map of the world, built and maintained by a community of volunteers. It holds detailed geographic data on roads, buildings, land use, points of interest, and much more, which you can download as a full planet file or as smaller regional extracts.
Overture Maps Foundation
An open map data project run under the Linux Foundation by a group of industry members. It combines and harmonises data from OpenStreetMap, Microsoft, Meta, and others into consistent, interoperable map layers you can build on freely.