UrbanOS-PoC Real-Time, Sovereign, Self-Heailing AI for City Mobility

Small update for anyone following UrbanOS. I refreshed the client templates and the MQTT simulations so implementation and testing are more straightforward. The client repo documents the MQTT publish contract, a Termux quick start, and DB field notes for map friendly ST_AsGeoJSON selects. The simulations repo includes load and capacity math that translates short runs to real fleet equivalents. UrbanOS is the intelligence layer, deployers own filtering and exposure. No telemetry, no cloud lock in, deployers own their data, the community decides what gets built on top. Happy to answer technical questions.

Links:

mqtt-client-templates: https://github.com/pablo-chacon/mqtt-client-templates

mqtt-simulations: https://github.com/pablo-chacon/mqtt-simulations/tree/main

UrbanOS-POC: https://github.com/pablo-chacon/UrbanOS-POC

Less than a day after I released UrbanOS someone had already cloned it and started experimenting with it. That was a very humbling moment for me. My only hope is that the project spreads and becomes useful for people. The goal of UrbanOS has always been to remain free and serve humanity, so every time I see it picked up or adapted it feels like confirmation of that vision.

For anyone interested, here's the link to the clone repo: UrbanOS PoC NYC adoption: https://github.com/jjiburg/UrbanOS-POC-NYC

I just released UrbanOS-PoC, a fully open-source system that uses real-time data to optimize public transport flow, without surveillance, black-box models, or cloud lock-in.

It runs on general-purpose hardware (Docker + Postgres + Python), uses MQTT for live data ingestion, and processes anonymized geospatial behavior to produce real-time routing and schedule predictions, without needing any personal identifiers.

I built this system solo over the last few years while raising two kids. I kept it closed for a while, hoping it might help our situation, but since it doesn’t rely on buzzwords, the attention was limited. So now it’s free for anyone who wants to work on urban planning, public mobility, or smart cities, without compromising privacy.

No prompts No user tracking No centralized platform Just logic, movement, and open AI for cities

You can simulate live data, inspect all modules, and spin it up locally on a laptop. Whitepaper, API docs, and demo setup included.

Repo: https://github.com/pablo-chacon/UrbanOS-POC Whitepaper: GitHub Wiki