BioRegional OS
A local-first operating system for a bioregional commons.
Real ecological boundaries in 3D, the BioRegional Commons protocol enforced in
code, and an AI you can point at any of it, running entirely on one computer,
with no account and no cloud.
Built by Green Fire Research and Innovation Center as part of its
bioregional AI and bioregional technology work, alongside
tools for eco villages,
permaculture and land projects.
More on what we mean by solarpunk technology.
Getting it
You need the folder on your computer first. Either:
- Download the zip from the green Code button at the top of
this page → Download ZIP,
then double-click it to unpack. No git, no account, no command line.
- Or, if you use git:
git clone https://github.com/Greenfire-solutions/Bioregional-OS.git
You also need Node, the big green LTS button. The Mac
launcher below checks for it and tells you the same thing if it is missing.
Start in one step
Mac: open the unpacked folder and double-click BioRegional OS (the icon).
It starts the OS if it is not running and opens it in your browser; if it is
already running, it just opens the browser. Click it as often as you like.
The first time, macOS may say it cannot check the app: right-click it, choose
Open, then Open again. That happens once. The first run also needs a few
minutes to set up, and does that in a Terminal window so you can watch it.
bin/Start BioRegional OS.command does the same job with a visible window,
if you would rather see what it is doing.
Anything else:
npm run setup
npm run os -- --open
Lost at any point? npm run help, or click the green I'm lost button in the
app. Broken? npm run doctor -- --fix.
New to this entirely → docs/START_HERE.md
What you see first is demonstration data, a fictional commons. A fresh
install opens on Barton Creek in Austin, Texas: an invented chapter, invented
people and invented observations, used to show how the twelve stages fit
together. None of it is a real report, and no organization named in it has
anything to do with it. Click Find my bioregion to replace it with your own
place, or npm run seed -- --reset to start over.
Live ecological data, water, weather, soil, species, hazards, is exactly that,
and comes from the public sources listed in
docs/DATA_SOURCES.md. The seeded commons around it
is fiction; the land it sits on is real.
What it does
A 3D map of the actual ecoregions you live in. Not decoration, live polygons
from the EPA Ecoregions Level III & IV dataset, extruded and interactive, with a
globe view for the wider mosaic. Your places, hubs and observations sit on top of
it. Every place resolves to its real ecoregion, biome and HUC12 watershed.
It starts where you are. The first thing a new person sees is one question,
where are you? answered live: your ecoregion, your watershed, the soil under
you, what lives around you, and the nearest gage measured against its own record.
Nothing is written down until you say so, and there is no account to make.
It tells you something before it asks you for anything. The first screen opens
with what the land is doing today: the nearest gage measured against its own median
for this calendar day across the whole period of record, live weather and any
official hazard alert, and sun, moon and daylight computed on your own machine, so
that part answers even with the wifi off. Then one line: what did you notice?
Every ecoregion in the country, readable with the network unplugged. npm run data downloads a dossier per region, plants and animals ranked by how often they
are actually seen, soil, climate normals, water, hazards, community assets, about
100 KB each, roughly 97 MB for all 967. Threatened species are named but never
located. Culture is deliberately not downloaded. Every run is resumable and
refetches only what has actually gone stale.
Live ecological data. USGS stream gages flow in as signals automatically,
discharge and gage height, refreshed on demand, cached so the map still works
with no internet.
Everything a person can do, Claude can do, and vice versa. One registry of
tools feeds the MCP server, the in-app assistant, the REST API and the forms in
the interface, the forms are generated from the tools' own schemas. There is no
privileged path that skips a consent gate.
A protocol that is actually enforced. The BioRegional Commons manual is not a
PDF sitting next to the software; it is the software. A council item without a
Land Seat report is refused. A project cannot reach the build stage with an open
consent gate. Paid work without acknowledged terms is refused. A chapter that
hasn't published what it does not represent fails its own viability test.
It tells you what came of it. When somebody's observation becomes a project,
and that project reaches the council, the person who noticed it is told so, by
name, on the first screen. Only observations a person made are ever attributed
to a person; a gage reading and a NOAA warning are credited to nobody.
It tells you what to do next. The Today view walks all twelve stages and
reports what is blocking, what has slipped past a date it committed to, and what
the protocol expects to exist and doesn't, each with the rule it comes from, so
you can argue with it, and the button that fixes it.
This is the beginning of the README. Read the rest at the original. Read the full README ↗
