The framing matters, so let us fix it first: Overture Maps is built on OpenStreetMap, not against it. Overture is a foundation project, founded in December 2022 by Amazon Web Services, Meta, Microsoft, and TomTom under the Linux Foundation, now with more than forty member organizations, that takes open map data, much of it OSM, and republishes it the way production software teams want to consume it: monthly releases, a fixed schema, stable identifiers, and cloud-native files. OpenStreetMap remains what it has always been: the living, editable source. This comparison explains what Overture adds on top of that source, where its data does and does not come from OSM, the licensing split that follows, and the practical rule for choosing between them. For the source itself, our open map database guide covers how OSM is built and licensed.
Home / GIS and Open Map Data for Beginners
OpenStreetMap vs Overture Maps: The Living Database vs the Production Snapshot Built on Top of It

Photo: Mikhail Nilov / Pexels
Living Database or Snapshot: Decide
Check off each step as you finish it. Your progress stays on this page.
0 of 7 complete
0% done
What Overture Adds: Schema, IDs, and a Release Train
OSM's strengths, anyone can edit anything, tags evolve organically, are also its integration costs: a pipeline consuming raw OSM must handle arbitrary tagging, shifting conventions, and features whose identifiers and geometry change as editors work. Overture's whole proposition is absorbing those costs once, centrally:
- A stable schema. Six themes, addresses, base, buildings, divisions, places, and transportation, each with a defined, versioned structure, so a query written this year keeps working next year.
- GERS, the Global Entity Reference System. Overture assigns stable IDs designed to persist across releases and to let different datasets reference the same real-world place, building, or road segment. Conflation, matching the same entity across datasets, is the unglamorous problem GERS exists to solve.
- Monthly releases as GeoParquet. Since July 2023 Overture has published monthly, distributed as GeoParquet files on Amazon S3 and Microsoft Azure storage, ready for engines like DuckDB, Spark, and BigQuery to query in place. No planet file to parse, no database to stand up before the first question.
- Scale. The releases carry on the order of 2.6 billion building footprints and more than 70 million places, numbers that tell you the intended user is a data team, not a person with a GPS.
In short, OSM is a database you can edit; Overture is a dataset you can depend on. Those are different virtues, and serious users often need both.
Where the Data Comes From, Theme by Theme
Overture is a conflation project: it merges sources per theme and says so. The buildings, divisions, transportation, and base themes incorporate OpenStreetMap data heavily, combined with other open sources such as satellite-derived building footprints and official boundary data. The places theme is the exception worth memorizing: it contains no OpenStreetMap data, being built instead from place data contributed by Meta and Microsoft and community sources such as AllThePlaces (which includes Foursquare-derived data). Addresses draw on open government address data in the countries that publish it. This per-theme sourcing is not a footnote; it determines the license you are under, which is the next section, and it means "Overture data" is not one thing with one rulebook.
The Licensing Split
Because the sources differ, the licenses differ, and this is the section to read before shipping anything:
- Themes containing OSM data carry the ODbL. The Open Database License's duties travel with the data: attribution, and share-alike on derived databases you publish. Overture does not launder OSM into a permissive license by republishing it.
- The places theme is published under CDLA Permissive 2.0, a permissive license without share-alike, reflecting its non-OSM sources.
The practical consequence: an application built only on Overture places faces a light license load, while one built on Overture buildings or transportation inherits the same ODbL obligations as using OSM directly. Anyone who tells you Overture is "OSM without the license" has read the places theme's terms and stopped. Read the theme-level license for the data you actually use, which is the same discipline our OpenStreetMap vs CARTO comparison applies to basemaps built on OSM.

Living Database vs Monthly Snapshot: The Working Differences
| OpenStreetMap | Overture Maps | |
|---|---|---|
| Freshness | Updated every minute; edits live immediately | Monthly releases; a snapshot as of the cut |
| Editable | Yes, by anyone, including you | No; corrections flow back through sources |
| Schema | Organic tagging, maximally expressive | Fixed, versioned, six themes |
| IDs | OSM IDs, which change as features are split and merged | GERS stable IDs, designed to persist |
| Format | Planet file, extracts, APIs, tiles | GeoParquet on S3 and Azure |
| Best at | Local detail, niche features, currency | Production pipelines, joins across datasets, repeatability |
The freshness row cuts both ways. A road that opened yesterday is in OSM today and in Overture next month; a pipeline that must produce identical results in March and September prefers Overture's frozen releases precisely because the world stops moving. Neither property is a defect. They are the same trade off as a news feed versus an archive edition.
Getting the Data: Ten Minutes to Your First Query
Overture's distribution is deliberately boring, which is a compliment. Each monthly release publishes its six themes as GeoParquet files in public cloud storage, on Amazon S3 and on Microsoft Azure, organized by theme, with release notes describing what changed. There is no account, no API key, and no tile server; you point a query engine at the files where they sit. Cloud-native engines and local tools that speak GeoParquet, DuckDB being the common example, can filter a theme by a bounding box and return just your city's buildings or places without downloading the planet, and the same query runs again next month against the new release with the schema unchanged, which is the entire point of the exercise. Budget your expectations by theme size: places is a comfortable download, buildings is a serious one, and the bounding-box-first workflow exists precisely so that seriousness is optional. Compare that with the OpenStreetMap path, a planet file or regional extract, a parser, a database import, covered in our OpenStreetMap download guide, and you have the two philosophies in one paragraph: OSM optimizes for a living database you host; Overture optimizes for files your existing data stack can interrogate today. One habit separates successful Overture users from frustrated ones: pin your pipeline to a named release and read its notes before upgrading it. The monthly cadence is a gift only if upgrades are deliberate; a pipeline that silently floats to the newest release has re-imported the instability the snapshot model exists to remove, one month at a time.
What Overture Is Not: No Tiles, No Router, No Editor
Knowing the boundaries prevents the standard mis-evaluations. Overture publishes no map tiles: there is nothing to drop into Leaflet as a basemap, because Overture is data files for pipelines, and rendering them is your stack's job. It runs no routing engine and no geocoder API; the transportation theme is road data your own tools must turn into routes. It accepts no edits: when Overture's places data is wrong, the correction path runs through the source (the contributing dataset, or OSM itself for OSM-derived themes), and arrives in a later monthly release, which is precisely the latency the snapshot model accepts in exchange for stability. And it is not a live service to query per request; you download release files, or query them where they sit on S3 or Azure with a cloud engine, and your application carries its own copy. Evaluate Overture as what it is, a well-engineered data distribution, and it is excellent. Evaluate it as a maps platform and you will be disappointed by absences that were never promises.
Verdict: Which One, When
- Choose OpenStreetMap when you need the freshest data, rich local detail and tagging (trails, benches, bike infrastructure, the long tail only contributors record), the ability to fix the map yourself, or data for rendering and routing today.
- Choose Overture when you are building a production data pipeline: you need a schema that will not drift, IDs that survive from release to release so you can join your own data against places and buildings, and files your warehouse or DuckDB can query directly, monthly.
- Use both, the pattern the architecture invites: Overture as the stable backbone for places and buildings in your pipeline, OSM as the living layer for everything current, local, and niche, with GERS and OSM IDs as the bridge. Overture's founders, several of whom are among OSM's largest users, built it to be consumed exactly this way.
Related Guides
The closest companions to this guide are the GIS and open map data guides below. Each one picks up a thread this page starts:
- Overpass Turbo: Query OpenStreetMap Like a Database, From Wizard to Export: Overpass Turbo turns OpenStreetMap into an answerable question: every bench in a town, every drinking water point along a trail, every bakery in a...
- QGIS Clip Raster: Cut Imagery and DEMs to Your Boundary With a Mask Layer: Rasters arrive as rectangles and study areas are never rectangles, so clipping is the first real operation most GIS work performs on imagery and...
- QGIS Elevation Profile: Read the Terrain Along Any Line You Draw: The Elevation Profile panel in QGIS turns a drawn line into a terrain cross-section: distance along the bottom, elevation up the side, with the...
Frequently Asked Questions
Is Overture Maps a replacement for OpenStreetMap?
No. Overture is built largely on OpenStreetMap data, republished in a production-friendly form: monthly GeoParquet releases, a stable schema, and stable GERS IDs. OpenStreetMap remains the living, editable source, updated every minute. Overture is a distribution of open map data, not a new map.
Who is behind Overture Maps?
Overture was founded in December 2022 by Amazon Web Services, Meta, Microsoft, and TomTom, as a project under the Linux Foundation (through the Joint Development Foundation). It has since grown to more than forty member organizations, including many companies that build on map data.
What is GERS?
The Global Entity Reference System is Overture's stable identifier system. Its purpose is to give real-world entities, a place, a building, a road segment, an ID that persists across monthly releases and can be shared across datasets, so different data products can be joined against the same entity. Raw OpenStreetMap IDs do not offer that stability, because features are routinely split and merged by editors.
Is Overture data under the same license as OpenStreetMap?
Partly. Themes that incorporate OpenStreetMap data, including buildings and transportation, carry the ODbL with its attribution and share-alike duties. The places theme contains no OSM data and is published under the permissive CDLA Permissive 2.0. Check the license of the specific theme you use.
Should I use Overture or OpenStreetMap for my app?
For rendering, routing, local detail, and freshness, use OpenStreetMap. For analytics and production pipelines that need a fixed schema, stable IDs, and repeatable monthly snapshots in GeoParquet, use Overture. Many production systems use Overture as the backbone and OSM as the living layer on top.
Photo credits: Mikhail Nilov (Pexels); Zelch Csaba (Pexels). Photos are served from this site; follow the credit to the library page for the original.