alt Hacker NewsShow HN: Offline tiles and routing and geocoding in one Docker Compose stack
Hi HN,
I’m building Corviont, a self-hosted offline maps appliance (tiles + routing + search) for edge/on-prem devices.
Hosted demo (no install): https://demo.corviont.com/
Self-host (Docker Compose repo): https://github.com/corviont/monaco-demo
Docs: https://www.corviont.com/docs
What’s inside:
- Vector tiles served locally (PMTiles)
- Routing served locally (Valhalla)
- Offline geocoding/search + reverse (SQLite Nominatim-based index)
- MapLibre UI wired to the local endpoints
Next (if people need it): a signed on-device updater for regional datasets (verify → atomic swap → reload).
I’d love feedback: where offline maps/routing/search matters for you, and what constraints bite (hardware, fleet size, update windows, regions, deployment style).
Comments
I like how you packed only the necessities - tiles for maps, routing, and geocoding index in in sqlite. I checked the monaco deployment and missed lookup with street number, as someone else also pointed out.
Why not create a "builder" repo, where people could generate their own local datasets by a bounding box?
I love it, thanks!
If I may have a feature request, I’d like to have only some of the features turned on - in my case it would be just the reverse geocoder (so I could skip the map and routing data download and storage).
Right now I have my own reverse geocoder for https://weathergraph.app which downloads OSM dumps and builds in-memory KD tree for lookups. Surprisingly, the whole world can fit in 3-4 GB of RAM, and service starts in 90 seconds on a cheap VPS, no database needed, but of course, having a battle tested solution that just works (and someone else maintains it) would help.
This is super cool. I’ve been kicking around an idea for ages regarding tile-based routing that I think would be excellent for offline routing. You could leverage the quadtree aspect of tiling to encapsulate faster, direct routes (ie highways) and as you go to deeper zoom levels you’d unlock small roads - even down to pathways. This keeps your in-memory graph small while traversing large distances (which would just be highways anyways) and once you eliminated most of the distance your remaining graph traversal on local roads would be small
Quite cool to see this space being explored! https://github.com/headwaymaps/headway is another related project.
Does this not require a massive database of tiles?
I ask because I've been looking to self host some sort of map tile server and they seem to have database in the hundreds of GB.
searching with street name and number offline would be nice. In google maps search only works when online
Definitely interesting, I don't see an obvious comment on hardware requirements, do you know what those are?
I've played around with OSRM, and Nominatim, etc, but had to do some trickery to run on a raspberry pi.
(For anyone interested in running some of these kind of things on a pi, I talk about it generally here, I need to post an update with more info at some point. http://blog.onaclovtech.com/2025/02/general-purpose-to-speci...)