Know which fields will cost you seed.
Our model reads each field's crop history from satellite images. Add soil samples where the risk is high. You contract the fields most likely to deliver.
See the planner

Our model reads each field's crop history from satellite images. Add soil samples where the risk is high. You contract the fields most likely to deliver.
See the planner

One reads soil for companies deciding what to do with their ground. The other reads it for biology nobody has found yet.
ESTABLISHEDSoil intelligence
Live with customers today
Erdaphon-1 combines our soil analysis, remote sensing and your own field data to find where your fields fall short. We test it on your land, then you scale what works and we re-measure every season.
FRONTIERBioassets
Early stage, built with partners
We search soil biology for candidates for new biological products, from what nature already evolved to assets we design together.
N.U. AgrarNew analytical methods allow us, as an independent consulting and research company, to see crop production from new perspectives. We uncover relationships that were once hidden and build the basis for ongoing improvements in sustainable arable farming. The soil’s microbial world is fascinating, and complex. Understanding it and turning insights into actionable, reliable recommendations requires a strong data basis. Reliable data collection and its clear presentation for accurate evaluation are the reasons why we work with Soilytix.
Lindhorst GruppeFor my site-specific management, the Soilytix analyses are especially valuable: they show how soil life and biological relationships differ between zones of a field, and help me manage each zone precisely. I learn about problems, differences and relationships I would otherwise never have known about.
Hof KöpkeThe analyses from Soilytix showed me which soils are best suited for potatoes. This allowed me to select my fields in a targeted way and position the crop optimally.
KWSWorking with Soilytix combines scientific depth with the responsiveness we expect from a trusted partner. They take complex soil questions and turn them into structured data we can act on across our pipeline.
Landwirtschaftskammer Schleswig-HolsteinSoilytix's analyses and expert guidance help us rigorously evaluate complex questions around sustainability and the soil impact of different nursery production systems, giving us the insights we need to drive the sector forward.
Maschinenring EHWWith the Soilytix software, we can organize sampling much more efficiently. All steps are digitally documented, the data flows directly into the system without detours, and we save ourselves a lot of rework.
Finck Stiftung / Gut & BöselThe Soilytix analyses provide us with solid scientific insights into the impact of our regenerative agriculture practices. We conducted a microbiome analysis in a diverse agroforestry system and, based on the composition of the bacterial, fungal, and microfauna communities, can draw conclusions about the influence of the trees on soil health and resilience.
Game Conservancy DeutschlandI really enjoy working with Soilytix, both privately with my own institute and through the Game Conservancy Deutschland. First of all, we receive state-of-the-art support, both in terms of data analysis and, of course, laboratory work. It's absolutely outstanding, the personal service is always available, no matter the time of day. We are extremely satisfied with the collaboration and the support that Soilytix provides us virtually around the clock.
YphenI'm a research and development manager at Yphen, a French startup that develops biosolutions for soil remediation. We started working with Soilytix to characterize the metagenomic profile and biological functions of our soil, particularly to identify species that may be able to remediate it. Needless to say, Soilytix’s dashboard, and especially their customer support, has been excellent. We’re really looking forward to deepening our collaboration and building an even better pipeline for soil analysis.
LWG BayernAs part of our research at the Bavarian State Institute for Viticulture and Horticulture, we have been working with Soilytix for some time. The comprehensive laboratory analysis they provide enables us to investigate complex questions within the Competence Center for Organic Horticulture project. With Soilytix, we can assess the effects of soil cultivation in vegetable production and the influence of cover crops on soil organisms. The detailed data preparation and the personal support we receive help us gain valuable insights for our work.
Recognised by![]()
One reference dataset. Your field data. Erdaphon-1 translates both into an indicator or recommender that shows where each field can do better.
Your fields, management history, your yields and trials, our soil analysis and satellite imagery.
Erdaphon-1
Tells you when to put nitrogen on each field, so the fertiliser you already buy feeds the crop rather than the soil life competing with it.
Ranks candidate fields for a given seed lot, so multiplication goes where the soil-borne risk is lowest and lots come back clean.
Flags which fields carry more pathogen pressure than comparable ground, so you can move the crop or the treatment before the season commits you.
Scores your land against comparable sites and shows which way it is moving, so you can prove the direction to a funder, a buyer or a board.
A team of scientists and market experts.
See the engineA team of scientists and market experts.
See the engineFive years, thousands of soil samples, proven protocols and more than 700,000 taxonomic references. Metabarcoding (16S, 18S, ITS) and metagenomics, physicochemical analysis, enzyme assays and more, all run to one protocol and enriched with remote sensing. INSTAND quality assessed, registered plant quarantine station.
PUBLISHED AND PEER REVIEWED
Schaks et al. A published microbial signature predicting plant growth in five independent datasets. Science of the Total Environment, 2025.
Read the paperLOAM reads the DNA of soil microbes the way a language model reads text. Four models, from 25 million to 624 million parameters, trained on public long-read genomes from soil and other environments. Ahead of comparably sized models on public benchmarks and competitive with much larger ones, LOAM turns raw DNA sequence into biological representations that help partners find and develop new solutions faster.
Model releases, method notes and field findings from our lab in Hamburg and our research team.

Research · 1 October 2026
A family of genomic language models trained on environmental DNA

Article · 31 August 2026

Method note · 20 August 2026
Tell us where you sit and what you need. We will reply as soon as we can.