Research, governance, and open collaboration for environmental intelligence.
The Naturecode Project is a nonprofit association dedicated to open research on AI for mapping, monitoring, and understanding the living world — from oceans and forests to soils, atmosphere, and the view from orbit.
Open research on AI for the living world.
We study how AI, Earth observation, and open data can make ecosystems more legible — across oceans, freshwater, forests, soils, atmosphere, and orbital observation — and how that legibility can serve conservation, food security, and the communities and institutions working closest to the ground.
Eight themes anchor our work.
Mapping nature
Earth observation across oceans, rivers, forests, soils, atmosphere, and the view from orbit — assembled into open, comparable layers of the living world.
Monitoring ecosystems in real time
Real-time signals, time-series, indicators, and change detection — so loss, recovery, and trend become legible at the cadence decisions actually need.
Analysis, prediction, and autonomous systems
Multimodal models, ecological digital twins, and agentic workflows — careful AI that knows what it does not know, and acts only when it should.
Valuing nature, biodiversity, and community work
How ecological evidence becomes a number a system can act on — for biodiversity, ecosystem services, and the communities doing the work on the ground.
Conservation, water, and food security
Translating measurement into decisions — for protected areas, restoration, fisheries, watersheds, agriculture, and the people working in each.
Efficient and sovereign computation
Smaller-footprint AI and edge-first, decentralized infrastructure — so the systems that read the planet do not cost the planet to run, and the data stays close to the people who produced it.
Open data and open collaboration
Open methods, open datasets, and shared infrastructure — built with universities, researchers, and the communities closest to the ground.
Governance, ethics, and data rights
How communities keep ownership of what they observe, how AI decisions are scrutinised, and how findings stay auditable — so the systems that read the living world remain accountable to the people who live in it.
Open research, in specific places.
Our themes meet the world in particular ecosystems, with particular partners. We work openly with the people closest to each place.
See all projects →Smart farming at Lumboakandhoo
Open research on island-scale smart farming in the Maldives — using AI to monitor poultry, soil, and water across a working farm, in partnership with Big Fish Maldives at Lumboakandhoo.
Mangrove conservation and monitoring in the Maldives
Open research on the status, change, and stewardship of Maldivian mangroves — combining satellite imagery, drone surveys, and custom AI systems with field observation and the knowledge of communities living alongside them.
Writing, published openly.
Briefs, notes, and longer-form papers from the project. We publish openly so that researchers, universities, NGOs, and institutions can adopt and challenge what we publish.
See all publications →Mapping the living world — a research agenda
Why the question "what is happening to the living world" has to be asked across oceans, rivers, forests, soils, atmosphere, and orbit at once — and what that means for AI, open data, and the communities and researchers we work with.
Refusal-first agents for ecological evidence
AI systems that decline to answer when the evidence is thin are more useful than systems that always answer. A short argument for refusal as a first-class behavior in ecological AI.
Multimodal evidence fusion for biodiversity monitoring
Satellite imagery, bioacoustics, environmental DNA, in-situ telemetry, and field notes are usually studied apart. We argue for studying them together — and outline what an open, multimodal monitoring stack would look like.
Open work, shared widely.
We work with universities, individual researchers, institutions, and the communities closest to the ecosystems we study. Different paths in — the same shared body of research.
Co-author, share methods, train together
Joint research on our themes, shared benchmarks, reproducible methods, and co-authored briefs and papers. We welcome PhD students, postdocs, and labs working in ecology, computer vision, remote sensing, and computational ecology.
Build evidence pipelines that hold up
Working with environmental agencies, parks, fisheries bodies, food-system programs, and conservation NGOs on open monitoring tools and ecological indicators that practitioners can actually use.
Ground truth from the people closest to it
Rangers, fishers, farmers, pastoralists, Indigenous and local communities — the people who see what satellites cannot. We collaborate without asking communities to surrender data rights to participate.