- On-premises vector tile maps
- Offline Mapbox Studio
- Mapbox APIs behind your firewall
Mapbox Atlas allows you to run Mapbox global maps on your own infrastructure, including behind a firewall or completely offline. Atlas includes an offline instance of the Mapbox Studio style editor, vector tiles, and has built-in support for custom tilesets. This means that you can run the Mapbox Maps APIs and Studio on a specific server or virtual private cloud that you have complete control over, then deploy to your web, mobile, or AR app.
Atlas supports the latest Mapbox vector map technology, including Mapbox GL JS, so maps render quickly and are highly customizable.
To install and use Atlas, you must have a Mapbox user account with a valid Atlas license. Atlas ships with a single-user account that can be shared concurrently among many users.
Need an Atlas license? Contact us to learn more.
Once you have added an active Atlas license to your Mapbox user account you will need to create a Mapbox token with
atlas:read scope within the same Mapbox user account.
- Go to the Mapbox access token page.
- Click on + Create a token.
- Give the new token a name like "Atlas Server".
- Uncheck all the default public scopes.
- Under Secret scopes, check the Atlas:Read scope.
- Click Create token.
- Save your new Atlas token in a secure place. Once you leave the Access Tokens page, you will not have access to this token again. Do not commit the Atlas token to GitHub or share it with anyone outside of your organization.
Below are approximate sizes for each dataset. Exact file sizes are subject to change with new data from periodic updates.
|Terrain-v2||1 GB||40 GB||820 GB|
|Streets-v7||300 MB||125 GB||Not Available|
|Streets-v8||300 MB||250 GB||Not Available|
|Countries-v1||1 MB||1.5 GB||Not Available|
|Satellite||1 GB||225 GB||2 TB|
|Terrain-RGB||1.5 GB||450 GB||3.1 TB|
|Boundaries||Not Available||15 GB||Not Available|
|Search||5 GB||75 GB||Not Available|
|Navigation||100 MB||50 GB||Not Available|
|Total||9 GB||1.2 TB||6.5 TB|
We recommend, as a minimum, the following hardware configurations. The development configuration is appropriate for testing and evaluating Atlas, and is not recommended for production workloads.
|2 CPU||4 CPU|
|8 GB RAM||16 GB RAM|
|60 GB SSD||2 TB SSD|