AGRIFY

Farmers illustration
AI Pepper
Field trials

Inspiration

Most farmers grow great produce but can’t prove it.

And when you can’t prove what happened on the farm, you get priced like a risk. No premium buyers. Slower export deals. And you’re basically invisible to financial products like export loans and insurance.

At the same time, buyers and regulators are asking for evidence, not paperwork. That’s the gap we built Agrify to close.

What we built

Agrify is a farmer first traceability and trade tool that turns everyday farm activity into verified records that buyers, exporters, lenders, and insurers can trust.

Instead of long forms, Agrify fits into real routines. Farmers speak quick updates, upload photos when they can, and pin their farm location once. Agrify does the hard work in the background, structuring the data and validating it so it’s usable for real decisions.

How it works

Farmer onboarding: Farmers create an account, share their farm location, and enter basic crop and season details.
Voice and photo capture: During land prep, planting, input application, and harvest, farmers log updates by voice or short text. They can add photos as proof. They do this just by talking.
Validation with Gemini plus geospatial context: We use Gemini to convert voice notes and images into clean, structured records. Then we validate submissions using crop stage expectations, image understanding, and location context (including metadata checks) to reduce fraud and increase confidence.
Produce Passport output: Agrify compiles verified activity into a Produce Passport that can be shared with buyers and partners, or displayed via QR code for audits and due diligence. This record can also be anchored on the XRPL for audits by partners.

Why this matters

For farmers, credible records become a trust signal they can actually use.

It helps them access premium offtakers, move faster through export checks, and unlock financial products like insurance and export loans because risk becomes measurable.

For buyers, it reduces sourcing risk and makes sustainability and Scope 3 reporting easier to support with evidence from the farm.

Challenges we faced

The hardest part wasn’t the AI. It was designing for real farmer behavior.

Connectivity can be unreliable. Time is limited. Documentation feels like extra work unless it clearly leads to money and market access.

So we focused on voice first capture, lightweight flows, and validation that increases trust without slowing farmers down.

What we learned

From tests with farmers, If documentation feels like a burden, it won’t happen.

But if it feels like a shortcut to better prices, faster trade, and access to finance, adoption comes quickly.

The real win is lowering friction while raising trust.

Built with

Google Gemini Speech to text for voice capture Image understanding for photo validation Geospatial context and metadata checks (GPS and EXIF) Web app frontend (React or Next.js) Backend API (Node.js or Python) Database (PostgreSQL or Firebase) Cloud storage for images and Produce Passport files QR codes for sharing Produce Passports Optional on chain anchoring for auditability (XRPL Produce Passport support)