Awesome Open Source
Awesome Open Source

A no nonsense Vector Tile pipeline

The purpose of Tiler is to create an easy to use, command line orientied pipeline for taking vector data in formats such as Shapefiles, and transforming them into raw Vector Tiles and MBTiles files (if required).

Tiler exists as a Docker container that unifies several technologies to streamline the creation of vector tiles.

CircleCI Badge

Blog Posts


Setup requires installation of Docker and a few Docker commands to get started. We've provided a nice little set of instructions in the SETUP file.

Using Tiler

Tiler provides a selection of scripts for converting between various formats and validating them (see the tiler-scripts folder). The primary and simplest way to use Tiler however is to use a config file placed in tiler/tiler-data/configs. Here is an example config file which references a shapefile and a table in a PostGIS database:


    "outdir" : "/tiler-data/tiles/",
    "tileset" : "states",
    "simplification" : 5,
    "validate" : false,
    "data" : {

        "states" : {
            "type" : "shapefile",
            "databaseInsert" : true,
            "paths" : ["/tiler-data/test-data/states/states.shp"],
            "minzoom" : 0,
            "maxzoom" : 5

        "capitals" : {
            "type" : "postgis",
            "query" : "select * from capitals"
            "minzoom" : 2,
            "maxzoom" : 10



This would be saved as tiler-data/configs/states.tiler.json. For full documentation of the tiler config, consult the CONFIG file in this repo. The file provides the location of the files you wish to translate, along with the output directory and if you want any simplification to occur. "data" is an object full of layers you wish to be ingested into the tiles. In this case the layer has multiple files that they use to generate that layer ("paths"). You can also provide a minimum zoom ("minzoom") and a maximum zoom ("maxzoom") for each layer.

At the moment "type" can be shapefile or geojson, postgis and gml.

Once your config file is setup, and you have followed the setup instructions you can run:

./ states

To generate the set of uncompressed vector tiles and an .mbtiles file.


Running Tests

A set of tests are provided that can be run using nosetest.

./ --test

Vector Tile Demo

A worked example; we could run the states config file as such:

./ states

After having ran this, there would be a set of states tiles in the tiler/tiler-data/tiles/states directory. You can then run a web server (i.e. python -m SimpleHTTPServer or live-server) from the host, with the tiler directory as the root. You can then navigate to the demos folder to try out those demo pages.

Demos are provided using Leaflet.VectorGrid and also Mapbox GL for you to display your tiles when you're done. You will need to specify your own styling in these examples.


See our contribution guide here Check out the roadmap here


This project was made possible thanks to building on a fantastic set of previous previous software:


MIT License

Get A Weekly Email With Trending Projects For These Topics
No Spam. Unsubscribe easily at any time.
python (53,198
dockerfile (871
converter (169
docker-container (112
geojson (81
vector-tiles (32
tile (29
shapefile (18