GitHub – opengeos/GeoLibre: A lightweight, cloud-native GIS platform for visualizing, exploring, and analyzing geospatial data. It runs in the web browser, on the desktop, on mobile, and inside Jupyter notebooks.

You can explore GeoLibre here: https://github.com/opengeos/GeoLibre

GeoLibre is one of those projects that feels practical before it feels flashy. That’s a good sign. It’s a lightweight, cloud-native GIS platform for visualizing, exploring, and analyzing geospatial data, and it runs in the browser, on desktop, on mobile, and inside Jupyter notebooks. The key detail is easy to miss at first, but it matters a lot, your data stays local and private.

That’s a nice fit for real work. Not the polished demo version of work, the actual version, where files come from different places, teams use different devices, and nobody wants to spend an afternoon wrangling formats just to inspect a map layer.

GeoLibre is built with Tauri v2, React, TypeScript, MapLibre GL JS, DuckDB-WASM Spatial, and deck.gl. In practice, that means you get a single workspace that can feel like a native app when you need it, then behave like a browser tool when that’s more convenient. I like that kind of flexibility. It cuts down on the usual “which app do I open for this?” problem.

The project also supports a surprisingly wide range of geospatial formats. GeoJSON, GeoParquet, GeoPackage, Shapefile, FlatGeobuf, KML, KMZ, GML, even GeoTIFF and COG rasters. You can drag files in, add layers, run DuckDB SQL directly in the browser, and use processing tools for vector and raster work. There’s even an AI segmentation workflow that turns imagery into vector features with SAM 3, although that part depends on a separate GPU-backed service.

For teams working in mixed environments, the deployment options are worth a look. GeoLibre can run locally, in Docker, or behind a subpath with the right proxy setup. It also includes options for offline or corporate setups, including mirrored runtime dependencies.

That combination is pretty telling. GeoLibre isn’t trying to be the loudest GIS tool in the room. It’s trying to be the one that fits into how people actually work, across devices, across environments, and without forcing your data into someone else’s system.

And honestly, that’s the part that tends to age well.