GeoLibre — A Free, Cloud-Native GIS for Every Platform • blog.dopana

GeoLibre is a free and open-source, cloud-native GIS platform — it runs on the web, on desktop (Windows/macOS/Linux), on Android, and right inside a Jupyter Notebook.

No server setup. No account required. Open a browser and start mapping.

Overview#

GeoLibre is built by opengeos ↗ (Q. Wu) with a modern tech stack:

Layer	Technology
Frontend	React + TypeScript
Map engine	MapLibre GL JS
3D rendering	deck.gl
Spatial SQL	DuckDB-WASM Spatial
Desktop & Mobile	Tauri v2
License	MIT

The key differentiator: all GIS processing runs in the browser — no backend, no separate database needed.

Where It Runs#

🌐 Web Browser#

Open https://geolibre.app ↗ — full featured, nothing to install.

💻 Desktop App#

Download the installer for Windows (.msi), macOS (.dmg), or Linux (.deb/.AppImage). The desktop app uses Tauri v2 — the bundle is just a few MB, 20x lighter than Electron.

📱 Android#

Native Android app on Google Play — full touch support, responsive across all screen sizes.

📓 Jupyter Notebook#

Python package that integrates GeoLibre directly into your notebook — for analysis and visualization workflows.

Tech Stack#

MapLibre GL JS#

The default map engine — supports 2D, 3D, vector/raster tile styling, basemaps from OpenFreeMap, Protomaps, EOX Sentinel-2 cloudless, and even planetary basemaps for the Moon, Mars, Mercury, Venus, Titan, Pluto…

DuckDB-WASM Spatial#

A spatial SQL engine running in the browser. GeoLibre uses DuckDB-WASM to:

Load local vector files (GeoJSON, GeoParquet, GeoPackage, Shapefile, KML, GPX…)
Run spatial SQL queries directly on your data
Client-side vector tiling for large layers
Export results to multiple formats

-- Run spatial SQL right in the browser
SELECT
  name,
  ST_Area(geom) AS area,
  ST_Centroid(geom) AS center
FROM loaded_layer
WHERE ST_Within(geom, ST_GeomFromText('POLYGON((...))'))
ORDER BY area DESC
LIMIT 100;

sql

deck.gl#

The 3D rendering engine for:

3D Tiles (Cesium 3D Tiles, ArcGIS I3S)
Point clouds, heatmaps, arc layers
3D building extrusion
Gaussian splats

Tauri v2#

The framework for desktop and mobile (replacing Electron):

Bundle size: a few MB (vs Electron’s ~200MB)
Rust backend for file system, native dialogs, system tray
Mobile support: iOS and Android

Key Features#

700+ Geoprocessing Tools#

All running in the browser — no server connection needed:

Buffer, clip, intersect, union, dissolve
Slope, aspect, hillshade, contour
Raster calculator, reclassify, zonal statistics
Network analysis, Voronoi, Delaunay
Spatial join, attribute join

SQL Workspace#

A built-in SQL editor powered by DuckDB-WASM Spatial:

Autocomplete for tables, columns, SQL keywords
Query history
Add results to the map or export
Combine DuckDB + PGlite (PostGIS in-browser) + Apache Sedona

3D & Planetary#

// Load 3D Tiles
const tileset = await add3DTiles({
  url: 'https://example.com/tileset.json',
  height: 100,
  color: [255, 0, 0, 200],
});

typescript

3D city models (NYC buildings colored by construction era)
Planetary basemaps: Moon, Mars, Mercury, Venus, Pluto…
Per-project ellipsoid — distance/area/scale measurements match the celestial body

Style & Symbology#

A powerful style engine compatible with QGIS and Mapbox:

Feature	Description
Symbology types	Single, categorized, graduated, rule-based, expression
Diagrams	Pie, donut, bar charts drawn on features
Labels	Data-defined label engine with full placement controls
Import/Export	OGC SLD, QGIS QML, Mapbox GL JSON
Color ramps	Inline preview gradient swatches
Heatmap	Point heatmap + clustering renderers
Raster styling	Pseudocolor, RGB band combination, colormaps

Legend Panel#

Auto-generated legend from visible layers:

Per-class rows for graduated, categorized, rule-based symbology
Gradient bars for heatmaps and raster colormaps
Proportional-symbol size ramps
Edit mode: rename, hide, reorder, collapse
Export JSON — saved with the project, shared with Print Layout

Data Integrations#

Connect directly to a wide range of data sources:

Cloud: Planetary Computer, Earth Engine, Overture Maps
Web services: WMS, WFS, WMTS, ArcGIS FeatureServer, OGC API
Local: GeoJSON, Shapefile, GeoPackage, KML, GPX, MBTiles, PMTiles
Raster: COG, GeoTIFF, NetCDF/HDF (kerchunk), Zarr
3D: 3D Tiles, I3S, glTF/GLB, Gaussian splats
CAD: DXF/DWG, Esri File Geodatabase (.gdb)

Multi-Map & Timelapse#

Multi-map grid: compare basemaps, layers, or time steps side by side
CesiumJS 3D globe in one pane (camera-synced with 2D maps)
Timelapse: animate annual cloudless basemaps (EOX Sentinel-2, NASA GIBS)

Wikipedia Knowledge Cards#

Click any place on the map — pull up its Wikipedia summary instantly.

Comparison with Traditional GIS#

	QGIS	GeoLibre	Google Earth Engine
Installation	~1GB	Web / ~5MB app	None
Server	Required	Not required	Required (Google)
Cloud-native	❌	✅	✅
In-browser	❌	✅	❌ (code editor)
Jupyter	Plugin	Built-in	API
Offline	✅	✅ (desktop)	❌
3D	Plugin	Built-in (deck.gl)	✅
Spatial SQL	❌	✅ (DuckDB + PGlite)	❌
Price	Free	Free (MIT)	Free tier
Mobile	❌	✅ (Android)	❌

Getting Started#

Web (nothing to install)#

Open https://geolibre.app → Start using it immediately

text

Desktop#

Download the installer from geolibre.app ↗:

# Or run from source
git clone https://github.com/opengeos/GeoLibre.git
cd GeoLibre
npm install
npm run desktop:dev

bash

Android#

Google Play → “GeoLibre” → Install

Jupyter#

pip install geolibre-jupyter