Mazorca: an offline AI agronomist for my cacao plot, powered by Gemma 4

Description and use

Mazorca is an offline AI agronomist for small cacao farms in Mexico, using Gemma 4 to diagnose pod diseases and produce weekly action plans.

Problem

Small cacao producers in Tabasco lack weekly technical visits and mobile signal, making it hard to detect and remove moniliasis (frosty pod rot) before it reduces harvest.

Outcome

Mazorca enables offline, confidence‑labeled diagnosis of cacao pod health, automated weekly reporting, and actionable plans without agrochemical recommendations, reducing reliance on technicians and improving data sharing for cooperatives and chocolate makers.

Workflow

• Walk the plot and take photos of pods and trees
• Upload photos to the local Mazorca web app
• Model generates a diagnosis card with findings and confidence levels
• Model creates a 7‑day report per lot and a weekly action plan
• Optional: Kestra flow builds reports on Monday, writes markdown, and alerts Slack if a lot worsened
• Export CSV for cooperatives or technicians
• Listen to the plan read aloud via ElevenLabs if desired

Evidence

Mazorca is a pocket agronomist that works offline. You walk the plot with your phone, take photos lot by lot, and an open‑weight model running on a laptop (Gemma 4 via Ollama, on a phone hotspot) tells you what it sees:

Source

It only suggests cultural practices: remove and bury sick pods, prune, manage shade, drain, harvest on time. It never recommends agrochemical doses; for any product it tells you to ask a technician.

Source

The farmer uses it for free. Whoever needs the data at scale pays for setup, support and the Kestra reporting.

Source