opengeos/GeoLibre

What it is

GeoLibre is an open-source GIS that runs in a web browser, as a desktop app, on Android and iOS, and inside Jupyter notebooks. It is aimed at researchers, students, instructors, and data scientists who need to inspect and analyze spatial data without maintaining a desktop GIS installation or uploading datasets to a hosted service.

The application reads cloud-native formats such as COG, GeoParquet, PMTiles, FlatGeobuf, Zarr, and STAC in place, and saves a map as a documented project file that others can reopen. GeoLibre is built with Tauri v2, React, TypeScript, MapLibre GL JS, DuckDB-WASM Spatial, and deck.gl. MapLibre is the default rendering engine; Mapbox GL JS, CesiumJS, and the ArcGIS Maps SDK can also be selected, including synchronized split panes.

Vector, raster, remote-sensing, hydrology, terrain, and LiDAR analysis is available through a toolbox of 1,000+ tools executed locally with WebAssembly. An optional, self-hosted Python sidecar can handle some tools. Planetary basemaps with per-project ellipsoids are available for non-Earth mapping.

Quick start

The fastest route needs no installation: open GeoLibre Web in a modern browser and use the processing and SQL tools there.

For local scripting, install the Python package:

pip install geolibre

or use conda install -c conda-forge geolibre. The Python guide describes extras such as geolibre[all] and how to embed the full app in Jupyter; a getting-started notebook is provided as a starting point. The R package is documented separately at r.geolibre.app.

To run the web application from source (Node.js 22+):

git clone https://github.com/opengeos/GeoLibre.git
cd GeoLibre
npm install
npm run dev

A desktop build additionally requires Rust. Docker and self-hosting instructions are in the getting-started guide.

Deployment and compatibility

Prebuilt installers are available for Windows, macOS, and Linux from the downloads page, along with distribution through the Microsoft Store, Mac App Store, Google Play, the App Store, AUR, and Flatpak. The browser version runs on any modern browser.

Analysis by default happens on the user's device, so processing performance depends on the local machine and browser. Some capabilities require external credentials: Mapbox GL JS and the ArcGIS Maps SDK need their own keys, as described in the rendering-engine guide. The desktop app, Android, and iOS builds are separate distribution targets with their own build documentation.

Testing and contributing

From a source checkout, the JavaScript suite runs with npm run test:frontend. Backend tests use a Python sidecar install with pip install -e "backend/geolibre_server[test]" followed by npm run test:backend. End-to-end tests use Playwright after npx playwright install chromium. npm run ci runs the full gate. Contributing, architecture, plugin API, and project-format references are linked from the README.

License

MIT. The project is archived on Zenodo with a citable DOI.

HotRepo's take

GeoLibre entered the all-language daily snapshot on 2026-07-28 at 3,145 stars with 420 stars gained that day, then recorded 667 and 671 daily gains on the next two days as total stars rose to 4,399. Its TypeScript daily snapshot on 2026-11-09 shows 8,506 stars and 66 stars gained, with a best position of 13 on that language list. The three consecutive all-language records show sustained attention during that July window, but only four snapshots are present and the final one is language-scoped and months later, so they support the observed counts rather than a continuous trend. The breadth of the deliverable — browser, desktop, mobile, Jupyter, R, and a 1,000+ tool WASM toolbox — is the notable technical fact for readers judging whether a local-first GIS fits their workflow.