From species lists to answers.

Researchers, conservation teams, watershed councils, and tribal and First Nations natural-resource programs are collecting eDNA at scale. They send a sample. They get back a list of species. They can't tell whether the result is healthy, recovering, or declining — because every lab runs its own assay, primer, taxonomy, and schema, and there's no standardized reference to compare against.

For a research lead, an agency program manager, or a natural-resource department, the practical effect is the same — every study is an island. Findings don't compose. Comparison across regions or years is manual, slow, and contested.

PHYLO sits downstream of any lab, assay, or sensor, so it's hardware-agnostic by design. Its agents pull in the relevant public data, harmonize preprocessing and taxonomy across institutions and schemas, and interpret results in context: what's there, how abundant it is, and how it's changing across sites, seasons, and years.

Where the work calls for it, we also build and deploy custom models trained on real eDNA, tuned to a specific ecosystem, assay, or question. Every deployment is shaped around the people using it. A single creek survey and a multi-year monitoring program get the same rigor, delivered at the scale each one needs. Indigenous data sovereignty (CARE/OCAP) and FAIR metadata are part of the design, not retrofits.

PHYLO's interpretation tools are already in use, from Chinook salmon nursery studies on Vashon Island to Indigenous communities in Indonesia. The end-to-end pipeline runs in production on real-world data, with results that hold up under expert review.

PHYLO runs across Google Cloud and AWS, supported by NVIDIA Inception, AWS Activate, and Google for Startups. The work has been validated by domain experts at the eDNA Collaborative, Conservation X Labs, and senior operators in commercial eDNA. We work with the research community, not against it. Our sister nonprofit, the Molecular Intelligence Institute, publishes the field's open standards, models, and interpretation — including under a perpetual license to this interface. Open substrate, proprietary platform.