How AgriTrust Scores Work
The engine turns raw wallet behavior into one explainable 0–100 score. Each of the five factors below can be inspected — so a farmer (or a lender) always knows why a number went up or down.
high trust
Example profile — 6 harvest seasons of verified activity
Factor breakdown
Weighted inputs to the displayed score
Core behavior metrics — season-to-season activity volume, counterparty count, and recency of logging.
How regular your farming activity is. Consecutive seasons without gaps raise this factor.
How recently you last logged activity. Fresh data means higher trust.
Total harvest value logged over time. Larger consistent volumes signal reliability.
Variety of counterparties and crop types. Diverse activity shows real operational breadth.
Verifiable, not trusted
Every score is shipped alongside a SHA-256 hash of the exact activity payload that produced it. Anyone can replay the inputs from public Horizon data and confirm the on-chain certificate matches what the engine actually saw.
Deterministic math, plain-English output
The score is computed deterministically — no AI in the loop for the number itself. An optional explanation layer translates the factors into language a farmer and a lender can both act on.