01
Detect
Where fires are, right now.
Active-fire hotspots are drawn from NASA FIRMS and Digital Earth Australia on a regular cycle, giving a current national picture of where fires have been detected.
Bushfire intelligence · Australia
Predict bushfire spread. Protect what matters.
Fire Path AI combines authoritative Australian data with established fire-behaviour models to estimate where a fire may spread - and when it could reach the assets you care about.
- Sources
- FIRMS · DEA · BOM
- Model
- McArthur + WindNinja
- Output
- Spread + time-to-impact
- Method
- Deterministic
Trusted data sources
- NASA FIRMS Active-fire hotspots
- Digital Earth Australia Satellite hotspots
- Bureau of Meteorology / Open-Meteo Weather + wind
- WindNinja Terrain-aware wind (optional)
- McArthur FFDI Fire danger index
By the numbers
A working pipeline, not a pitch deck.
- 5
- Data & model sources FIRMS · DEA · BOM · WindNinja · McArthur
- 10 min
- Fastest hotspot refresh Digital Earth Australia geostationary feed
- 1–12 h
- Modelled spread horizons a corridor that bends with the wind
- 4
- Stage pipeline detect · model · map · alert
Structural facts about how the platform is built — not performance, accuracy, or uptime claims.
The problem
Bushfire risk moves faster than the paperwork.
When a fire starts, the people responsible for homes, infrastructure, and land need to know one thing quickly: is it coming our way, and how long do we have? That answer depends on terrain, wind, fuel, and where the fire already is - data that exists, but rarely in one place, mapped against the specific assets that matter to you.
Fire Path AI brings those sources together and turns them into a clear, repeatable picture of modelled spread and estimated time-to-impact.
How it works
From a satellite detection to your doorstep
One deterministic pipeline. The same inputs always produce the same, inspectable result — so the live map and the warning that reaches you can never disagree.
Authoritative inputs
- FIRMS
- DEA
- BOM
- WindNinja
- McArthur
-
- 02
Model
Where a fire may spread, and when.
Established fire-behaviour models (McArthur) estimate spread from the detected fire, driven by weather and wind - refined over local terrain with WindNinja where terrain processing is enabled.
- 03
Map
Against the assets that matter to you.
Modelled spread and estimated time-to-impact are shown against the specific places you have registered - homes, sites, infrastructure, and land.
- 04
Alert
To the people responsible.
As risk to a registered asset changes, the people responsible for it are notified - so the right person acts on the right information in time.
Delivered to the right person — in time to act
Capabilities
What Fire Path AI does
Fire behaviour modelling
Wind-adjusted spread modelling on the McArthur Forest Fire Danger Index, with optional WindNinja terrain-aware wind fields where terrain processing is enabled.
Satellite hotspot detection
Active-fire hotspots from NASA FIRMS and Digital Earth Australia, mapped against your registered assets.
Asset protection
Register the places that matter and see modelled spread and estimated time-to-impact as conditions change.
Alerting
Configurable alerts as risk to a registered asset changes - delivered to the people responsible for it.
Organisations & roles
Organisations with role-based access, so teams see and manage only what they should.
Deterministic by design
Predictions come from published physics and rules-based models - transparent and repeatable, not a black box.
The prove-it loop
We measure the prediction — we don't just claim it.
Anyone can draw a cone on a map. The harder, honest question is whether it was right. Fire Path AI keeps score — the mechanism that turns a projection into evidence.
-
1 Predict
Each modelled spread corridor is stored immutably, stamped with the exact model version that produced it — so it can never be quietly rewritten after the fact.
-
2 Observe
As the fire moves, its real perimeter is recorded over time from the same satellite feeds that detected it.
-
3 Score
Prediction meets reality: overlap, head-position error in kilometres, and containment — did the fire stay inside the corridor we drew?
-
4 Refine
Scores accumulate per model version, so every change to the engine is evidenced against real events, not assumed.
Accuracy figures are deliberately not published yet — they require scores across real fire events and expert sign-off. The mechanism is built and running; the claim waits for the evidence.
Why deterministic
Transparent models, not a black box.
Fire Path AI v1 produces predictions from published physics and rules-based models. That is a deliberate choice for life-safety and government use.
- Repeatable. The same inputs produce the same output, every time.
- Inspectable. Results trace back to the data and models that produced them.
- Auditable. No opaque scoring standing between the inputs and the outcome.
On the roadmap: learning from historical fires to help refine spread prediction over time.
Who it's for
Built for the people responsible
Government
Deterministic, auditable predictions with Australian data residency - built for agencies responsible for people and land.
Learn moreEnterprise
Asset protection at scale across sites and teams, with role-based access and configurable alerting.
Learn moreCommunities
Designed with homeowners and local communities in mind as the platform grows.
Learn moreRequest a briefing
See Fire Path AI applied to your assets.
Request a briefing and we'll walk through the modelling, the data sources, and how it maps to what you protect.