OpenStreetMap is the Wikipedia of maps. It is a free, open, crowd-sourced map of the entire world, built by a community of millions of volunteers and used as the data backbone for thousands of products you already use, from weather apps to hiking apps to ride-share platforms.
This guide explains what OpenStreetMap actually is, how it works as a data set, how it differs from Google Maps, and what role it plays inside a modern production mapping stack like MapAtlas.
The Project, in One Paragraph
OpenStreetMap (OSM) was founded in 2004 in the United Kingdom by Steve Coast, after he grew frustrated with the lack of freely available digital map data. The idea was to build a map of the world the way Wikipedia was building an encyclopedia: anyone can sign up, anyone can add or edit a feature, and the resulting data is published under an open license. Today the project has more than 10 million registered editors and an active core of around 60,000 monthly contributors. The full planet database is over 100 GB compressed and grows every minute.
What OSM Looks Like as Data
OSM is not a folder of map tiles. It is a structured database of three primitive types:
- Nodes: single points with a latitude and longitude. Used for things that are essentially a point (a bench, a tree, a parking meter), and as the building blocks of more complex shapes.
- Ways: ordered sequences of nodes. Used for linear features (a road segment, a river) and closed shapes (a building footprint, a park boundary).
- Relations: groupings of nodes and ways. Used for things like bus routes (a sequence of road segments), administrative boundaries (a polygon made up of many ways), and multipolygon features.
Every node, way, and relation has a set of key-value tags. A road might have highway=residential, name=Rue de Rivoli, maxspeed=30. A coffee shop might have amenity=cafe, name=Cafe Central, opening_hours=Mo-Fr 07:00-19:00, wheelchair=yes. The tagging vocabulary is community-defined, evolving, and documented on the OSM wiki.
This combination of open primitives and tagged metadata is what makes OSM so reusable. You can extract every cycle path in Berlin, every defibrillator in the Netherlands, or every wine region in France with one structured query.
How the Map Gets Made
OSM is edited continuously by volunteers using web editors (iD), desktop editors (JOSM), and mobile editors (Vespucci, StreetComplete). Edits flow into a central database within minutes. Source material includes:
- Aerial imagery licensed for OSM use (Bing, Maxar, Esri)
- GPS traces uploaded by community members
- Government open data (cadastre, road inventories, address files)
- Local knowledge, especially for opening hours, names, accessibility, and small features
- Mass imports of structured data sets (carefully reviewed, since bad imports damage data quality)
Quality control is community-driven: experienced mappers review changes, automated tools flag suspicious edits, and per-region communities maintain local conventions. The result is a map that is unusually detailed in places where the community is active (Western Europe, urban North America, Japan) and patchier in less-mapped regions, where commercial alternatives sometimes have an edge.
OSM vs Google Maps
The most common question developers ask is whether to build on OSM or on Google Maps. The honest answer depends on what you are building.
OSM strengths: open data you can host, redistribute, and analyse; richer point-of-interest tagging in many regions; transparent provenance; full European data sovereignty; predictable pricing through OSM-based providers; the ability to render custom map styles for your brand.
Google Maps strengths: Street View imagery; very strong business listing data in some regions; consumer brand recognition.
For developer-grade products that need EU hosting, GDPR-clean data flows, custom styling, and predictable pricing, an OSM-based stack is usually the better choice. For consumer products that lean heavily on Google's business reviews or Street View imagery, Google has unique value.
How Production APIs Build on OSM
You almost never call the OSM database directly in production. Instead, you use a hosted API that builds on OSM data and adds the layers OSM does not provide on its own:
- Tile rendering: rendering 100 GB of vector data into fast map tiles for every zoom level requires a serious tile pipeline. Hosted providers do this for you.
- Geocoding: turning addresses into coordinates needs an indexed search engine on top of OSM, plus heuristics for ranking and disambiguation.
- Routing and isochrones: OSM has the road graph, but computing actual driving directions and travel-time polygons needs a routing engine (OSRM, Valhalla, GraphHopper) running on dedicated infrastructure.
- Map matching, distance matrices, optimisation: same story.
- SLAs, EU hosting, GDPR compliance: hosted providers package the data with the contract terms enterprise teams need.
MapAtlas is the European mapping platform built on OSM-grade data with EU-only hosting, GDPR-clean infrastructure, and developer-grade APIs for geocoding, routing, isochrones, distance matrices, and dynamic styled maps.
When You Need Direct OSM Data
Sometimes you do want the raw data. Common cases:
- Bulk analysis: counting all electric vehicle charging stations in Germany, mapping every defibrillator in Spain.
- Custom GIS workflows: importing into PostGIS for spatial joins.
- Internal map products: rendering your own tiles with full control over the style.
- Data enrichment: adding OSM tags to your own location records.
For these, the Geofabrik extracts or Overpass API are the standard entry points. Both are free, both run on volunteer infrastructure, and both come with attribution and license terms you should respect in any redistributed product.
Why MapAtlas
MapAtlas builds developer-grade APIs on top of OSM-derived data with EU-only hosting and GDPR-clean infrastructure. The Geocoding API, Search API, Directions API, and Isochrone API all return clean, well-tagged location records that work cleanly in production code and in the structured data you ship to AI assistants. For a hands-on intro, try the Coordinates Lookup tool or read What Is a Geocode for the deeper context.
Frequently Asked Questions
What is OpenStreetMap?
OpenStreetMap (OSM) is a free, open, crowd-sourced map of the entire world, maintained by a global community of volunteers. Anyone can edit the map, add new roads or places, and reuse the data under the Open Database License. Behind the scenes, OSM is a structured database of nodes, ways, and relations that describes streets, buildings, points of interest, transport networks, and natural features. Most of the maps you see on weather apps, hiking apps, ride-share apps, and developer-grade map APIs are powered by OSM data, often with proprietary enhancements layered on top.
Is OpenStreetMap free to use commercially?
Yes. OpenStreetMap data is licensed under the Open Database License (ODbL), which allows commercial use as long as you attribute OpenStreetMap and share derivative databases under the same license. The basic tile server at openstreetmap.org has acceptable-use limits for production traffic, so most teams use a hosted tile provider (MapTiler, Mapbox, MapAtlas, etc.) that builds on OSM data and offers SLAs, custom styles, and higher rate limits.
How is OpenStreetMap different from Google Maps?
OpenStreetMap is open data: anyone can download the entire planet and build their own product on top of it, and the editing community is fully transparent. Google Maps is a closed proprietary product: the underlying data is not redistributable, and pricing depends on Google. For most developer use cases (geocoding, routing, custom map styles, EU-hosted infrastructure, GDPR-clean data flows), an OSM-based provider gives more control and predictable pricing. For features that need Street View imagery or Google's specific business listing data, Google still has unique data.
What is an OSM tag and why does it matter?
Every feature in OpenStreetMap has a set of key-value tags that describe what it is. A road might be tagged `highway=residential`, a restaurant `amenity=restaurant`, a bus stop `highway=bus_stop`. The tagging vocabulary is community-defined and documented on the OSM wiki. Tags are what let downstream tools filter the data sensibly: you can query 'all hospitals in Paris' or 'all charging stations in Berlin' because each feature is tagged consistently. Production geocoders (including MapAtlas) lean on OSM tags to score the relevance of a match.
