AI agronomist · live on a Sonoma County farm

Stop spraying 80 acres for a 3-acre outbreak.

Farmers do not just need detection. They need memory.

A Sonoma strawberry grower told us he sprays far beyond the affected block because he cannot see whether pests are spreading. DrCrop turns scout reports today, and drone or field-camera evidence tomorrow, into GBrain memory across crop, pest, location, time, and conditions. Then the AI agronomist returns a biological-first plan: where to scout, what to treat, and which acres to leave alone. Fewer unnecessary sprays means lower residue risk on food.

Brief the agronomist Source on GitHub

Built on GStack + GBrain · context retrieval by ZeroEntropy · agronomist powered by Claude Sonnet 4.6 · written for Garry Tan's RFS for low-pesticide agriculture

Three steps. Every observation. Every block. One connected memory.

  1. 1

    An observation lands

    Human scout today. Drone, phone photo, or field camera tomorrow. All the same shape: crop, pest, zone, severity, time, symptoms.

  2. 2

    GBrain connects the dots

    Typed edges link the new report to nearby blocks, recent outbreaks, matching pest profiles, and the wind-aligned spread path.

  3. 3

    The AI agronomist recommends action

    Scout first. Use biological controls when they fit. Escalate to chemicals only if scouting confirms the threshold.

Live console

Farm memory

Sonoma County · scout it like your farm

Active outbreaks 0
Targeted acres 0
Spray acres avoided 0
Memory confidence 0%

Live blocks

Field map

Click a marker to load that observation into the action plan.

New scout report

Observation

Live AI briefing

Targeted action plan

Moderate

Press Generate briefing to ask the Claude-powered agronomist for a plain-English plan grounded in the memory graph.

Action plan

Recommendations

    GBrain shape

    Memory graph

    
          
    

          
    
            
    
              
    
                
    Reports
    
                
    Timeline