← All help

Supported file formats

Every format DropMaps can read, and how it handles them.

Vector

FormatNotes
Shapefile.shp plus sidecars, or a single .zip
File GeodatabaseEsri .gdb zipped as .zip / .gdb.zip — all feature classes read
KML / KMZPlacemarks, paths and polygons
GPXGPS tracks, routes and waypoints
GeoJSONFeatures and feature collections
TopoJSONConverted to GeoJSON on load
GeoPackage.gpkg — every vector layer read
FlatGeobuf.fgb streaming vector format
CSVPoint data from lat/lng columns, or geometry from a WKT column
GeoParquet.parquet / .geoparquet read via DuckDB + spatial

Raster

FormatNotes
GeoTIFFRead directly in the browser
Cloud-Optimized GeoTIFF (COG)Streamed via HTTP range reads — no full download
PMTiles (raster)Pre-tiled imagery served from a single archive
GeoPDF.pdf maps that carry georeferencing — see GeoPDF below

Tiles

FormatNotes
PMTiles (vector)Vector tiles from a single portable archive
MBTilesTiles from a single SQLite-based archive

DropMaps map files

FormatNotes
.dmapA whole DropMaps map — layers, styling and data in one portable file. Export and re-open it in the app, or move it to and from QGIS with the QGIS plugin.

Need to make a big GeoTIFF or MBTiles stream well? The Tile Converter turns them into COG and PMTiles locally, with no file-size limit.

GeoPDF

A GeoPDF (or “geospatial PDF”) is an ordinary PDF map that also carries its coordinates inside the file. Drop one on the map and DropMaps reads that georeferencing, renders the page to imagery and places it like any other raster layer.

Both of the encodings in circulation are read:

EncodingWritten by
OGC / ISO 32000 (the /VP viewport)QGIS, ArcGIS Pro, GDAL, most modern exporters
TerraGo (the older /LGIDict)USGS US Topo quads, Avenza-era map sheets

You don’t need to know which one you have — DropMaps tries both.

What you get

  • A picture, not data. This is the big one. A GeoPDF stores the map as drawn artwork, so there are no features to select, style, query or edit — you get an image layer, the same as a GeoTIFF. Roads in the PDF are lines someone drew, not a roads layer. To work with the underlying data, go back to the source it was published from.
  • Cropped to the map frame. The page is cut to the map’s neatline, so the title block, legend and margins around it don’t land on the map as a white box.
  • The main map only. If a page holds several map frames — an inset, a locator map — the largest one is imported and the smaller ones are ignored.
  • Named from the PDF. The layer takes the map frame’s name or description if the file has one, otherwise the filename.
  • Its coordinate system is recorded, so it shows in the layer’s Details panel and behaves like any other layer with a known CRS. See Coordinate systems.
  • It becomes a COG behind the scenes, which means it streams as you zoom, survives a page reload, and — on a paid plan — backs up and syncs to your other devices like any other raster.

Resolution

Pages are rasterised at 150 DPI, which keeps map lettering and contour labels readable. Because the map becomes pixels, it will soften if you zoom far past the scale it was drawn for — a 1:25,000 sheet looks the way a scan of it would.

Very large pages (big wall-sized plots) are rendered at a lower DPI so they fit in what a browser can hold in memory; on iPhones and iPads the ceiling is tighter again. When that happens you’ll see a note telling you what resolution was actually used.

Rotated maps

Most map sheets are north-up, and those are placed without altering a single pixel — the georeferencing is simply attached.

If the map is rotated on the page (a corridor or coastline sheet turned to fit), it has to be straightened to sit on a north-up web map, so it’s resampled once. Fine lettering softens very slightly, and you’ll see a note saying it happened. The area around a rotated map is left transparent, not filled in, so your basemap still shows through.

What isn’t read

  • Vector geometry. Nothing is extracted as features — see above.
  • PDF layers (OCGs) and scale-dependent content. The page is rendered as it looks when you open it, with whatever layers the author left switched on by default.
  • Attributes and attachments embedded by tools like Avenza or ArcGIS.
  • Other pages. If several pages are georeferenced, the first is imported and you’re told how many there were.

If it doesn’t work

What you seeWhat it means
”This PDF has no georeferencing”It’s a plain PDF — a map printed or exported without coordinates. DropMaps won’t guess at a position. Re-export it as a geospatial PDF, or georeference it in QGIS and export a GeoTIFF.
”a georeferencing dictionary DropMaps can’t read”A TerraGo file using a projection that can’t be decoded. Re-export from QGIS or ArcGIS Pro, or convert to GeoTIFF first.
”georeferenced from its corner coordinates instead”The file’s projection couldn’t be resolved, so the map was placed from its corners. Placement is very close, but not exact.
It lands in the wrong placeCheck the source: some PDFs carry georeferencing that is simply wrong. Open it in QGIS to confirm before assuming it’s the import.

Tip: if you need the actual data rather than a picture of it, look for the source layers first. Many agencies publish the same sheet as a shapefile, GeoPackage or Feature Service — importing that gives you something you can query, style and edit.

Projections

Layers are reprojected to WGS84 (EPSG:4326) automatically using proj4, so data in local or national grids lines up with everything else on the map. Your original CRS is recorded and can be reprojected back to on export.

See Coordinate systems & projections for how the original CRS is detected per format, and one limitation to know about: exporting “in the original CRS” is a round trip through WGS84, so it isn’t bit-for-bit identical to your source file.