Most wild places aren't lost in the field. They're lost on paper — in scoping notices that close before anyone responds, in environmental assessments that go uncommented, in permits that pass review without ecological challenge. By the time the bulldozers arrive, the decision is already final and the legal window for challenging it has long since closed.
The administrative record — the formal collection of comments, evidence, and analysis that agencies are legally required to consider — is where outcomes get fixed. It is also where conservation organizations are most absent. The federal land management agencies operating across Cascadia process thousands of decisions a year. Small ecological organizations participate in a tiny fraction of them. Most of those decisions go through without anyone challenging the assumptions, surfacing the cumulative impacts, or making the legal record that later litigation depends on.
This is not because the work is impossible. It is because the volume vastly exceeds the capacity of small, grassroots organizations to track, analyze, and respond.
The Conservation Defense Initiative is the infrastructure we are building to make ecodefense systematic, automated, and replicable.
The problem CDI addresses is operational: the throughput of administrative actions threatening ecologically significant landscapes vastly exceeds the human capacity of small, grassroots organizations to monitor and respond. Most threats pass without challenge — not because we lack the knowledge to intervene, but because we lack the throughput.
CDI is built on a stack of open and purpose-built tools: automated scrapers monitor public agency document repositories (USFS NEPA trackers, BLM NEPA registers, Oregon DLCD, California CEQA portals) and pull new filings as they are posted. Each document is parsed and classified using a combination of rule-based filters and a large language model — currently Claude — scored against ecological watch criteria and flagged when it crosses the threshold. For flagged actions, a retrieval-augmented drafting system grounded in our archive of prior interventions, applicable case law, species recovery plans, and regional land management documents generates first-draft comments. A structured database logs every detected action and its lifecycle, producing the outcome data needed to measure impact over time.
The intent is not to remove human judgment from ecodefense — every intervention passes through human review before it goes out. The intent is to eliminate the upstream labor cost that currently prevents small organizations from being present in the administrative record at all.
CDI is explicitly designed to be replicable. The architecture is geography-configurable; the knowledge base is structured for sharing; the documentation will be open. Other small ecodefense organizations facing the same throughput problem will be able to adapt and deploy it.
If your organization does ecodefense work and needs a fiscal home, we sponsor projects like this. Our sponsorship program provides 501(c)(3) status, financial administration, and organizational support to grassroots initiatives — including ecodefense work that struggles to find conventional fiscal sponsors.
Read about sponsorship →If you have legal or scientific expertise relevant to administrative ecodefense and would like to contribute to CDI's knowledge base or review interventions in progress, get in touch.
Contact us →If you want to support this work financially, donations directly fund ecodefense capacity — staff time, technical infrastructure, and the ongoing curation that makes the practice possible.
Support the work →