Today's grain storage tools detect spoilage weeks after biological activity begins. By then, mycotoxin accumulation may already trigger FDA grade downgrades or load rejection.
GrainGuard closes that gap.
A scanning MOX array continuously samples grain headspace chemistry, generating a
multi-channel mVOC fingerprint. A trained neural network classifies it in real time against known mold and pest signatures and sends mobile alerts - before spoilage is visible or detected by current technologies.
Validated performance (March 2026)
- Detection onset at day 9 - ~12 days before visible damage
- Zero false positives across ~270,000 baseline readings
- Cohen's d = 3.13
A provisional patent was filed December 2025. Two planned farm pilots are underway, one in South Carolina and one in Tennessee.
We're looking for:
- A partner for collaborative development to expand training to include the grain substrates, mold species, and stored-product insects of greatest relevance to on-farm storage operators and commercial grain elevators.
- Field trial partners for multi-commodity silo validation, and co-validation of mVOC onset against HPLC mycotoxin assay.