I'm a big fan of QGIS, but doing stuff browser side is just _terribly_ convenient (I also never got anything 3D to work on my M2 Mac with QGIS, but granted, haven't tried QGIS 4 yet) and I admit that I enjoy working with public datasets that are hosted on e.g. ArcGIS Online. https://share.geolibre.app is super neat, too.
It's funny to run into this particular project on HN: My side project (https://skaldmaps.com, if I may) uses a stack _very_ close to it, which is to say MapLibre + DuckDB + React + a PMTiles cache, and it's been super pleasant to work with (especially combined w/ dbt and a more "traditional" data stack). I also think it performs really well.
All that to say, definitely adding this one to the homelab!
QGIS will either get an LLM-assistent rewrite, or the next QGIS is already in the making in Rust/Swift somewhere. It is incredible it still works in v4.
Gave this a little go. The web version was giving me IO errors for any file I tried to load in. The desktop version worked a little better and could render some smaller files (~30mb geopackage, 300kb SHP.zip).
Trying to load some bigger files >1gb kind of just sat on 'Importing data...' for a while, then the screen went blank and I lost all my layers.
Got mixed results with a grab bag of other spatial files I had laying around. A lightweight alternative to QGIS would be good but it feels like this has a way to go.
This looks like an alternative to ArcGIS online Map Viewer. If so, this is exciting! A subscription free service for gathering data for non-profits in the field using web based tools.
Side note: one tell for AI speak is not understanding what’s important. This boasts that it adapts to mobile screens, which is hardly unusual for a modern website and probably not a central feature of the software.
Wow quite nice. I’ve been working on a side project (so all vibecoding) a tool for canyon mapping (Canyoneering). I made a decent interface based on Leaflet and react but I struggled to add a more complex interface without just recreating the enormous QGIS GTK 2000s era UI. Look forward to consider adopting this stack and UI instead of rolling my own.
Very cool to see. I've been vibecoding a side project as well for learning/using the cloud-native geospatial stack (https://sylveapp.com/, quite rough around the edges) that focuses solely on processing and hosting tiles. I'm still working on getting a usable demo for my example map but seeing this helps to refine what I'm working; to focus on hosting/publishing vs visualizing/analysis.
It's fun to see all other projects that also use a similar stack (MapLibre + DuckDB + PMTiles).
That seems like a really cool project. Has a lot more detail on some things than google maps - where trash cans are etc. Quick glance around my local hood & seems to check out
The data you're talking about is (assuming you're using the default basemap) OpenStreetMap's, a different project that pretty much powers everything not Esri or Google (and some of those too) these days.
If you know what you're doing, have a solid vision and plan then AI amplifies the clarity I guess but I doubt something this polished was made in 2 weeks :)
It will be interesting to see how well it adapts and avoids regressions through future releases. The usual downfall of this vibecoded stuff is long term sustainability.
I'm a big fan of QGIS, but doing stuff browser side is just _terribly_ convenient (I also never got anything 3D to work on my M2 Mac with QGIS, but granted, haven't tried QGIS 4 yet) and I admit that I enjoy working with public datasets that are hosted on e.g. ArcGIS Online. https://share.geolibre.app is super neat, too.
It's funny to run into this particular project on HN: My side project (https://skaldmaps.com, if I may) uses a stack _very_ close to it, which is to say MapLibre + DuckDB + React + a PMTiles cache, and it's been super pleasant to work with (especially combined w/ dbt and a more "traditional" data stack). I also think it performs really well.
All that to say, definitely adding this one to the homelab!
Trying to load some bigger files >1gb kind of just sat on 'Importing data...' for a while, then the screen went blank and I lost all my layers.
Got mixed results with a grab bag of other spatial files I had laying around. A lightweight alternative to QGIS would be good but it feels like this has a way to go.
It's fun to see all other projects that also use a similar stack (MapLibre + DuckDB + PMTiles).
It will be interesting to see how well it adapts and avoids regressions through future releases. The usual downfall of this vibecoded stuff is long term sustainability.