Croppie – the PhotoCropping app
AI enabled coffee yield estimation
Croppie is an AI-enabled mobile app that uses simple smartphone images to improve coffee yield estimates. Coffee yields vary widely due to management practices and climate variability, but reliable estimates are essential for financial planning, precision agronomy, and traceability. Traditional manual methods are slow, costly, and unreliable.
Croppie addresses this by guiding users through a simple field protocol to capture images of productive branches and basic plot information. The app automatically estimates yield, reducing labor and skill requirements while improving consistency and transparency.
Designed for smallholder systems, Croppie enables scalable, low-cost data collection that integrates seamlessly into digital advisory, finance, and value chain applications, supporting better decision-making across the sector.
In what context is this tool useful?
The Croppie app specifically targets small-scale farmers and field personnel who may have limited access to resources, information, and technology. It is most useful, when data provides added value to the crop value chain, e.g. for credits, sales or regulatory compliance in high value markets.
Variations, Scaling and Adaptations
Each development phase follows a three-step, human-centered development pathway that ensures technical robustness, coffee system relevance, and user acceptance. The approach combines agronomy, value chain analysis, human-centered design (HCD), and AI development.
The existing app is available in the Android app store, or visit:
Expected Results
Croppie AI assisted yield estimates demonstrated the potential to make coffee yield estimation less labor and skill intensive, easier to integrate in data-driven value addition, and thereby offering greater scaling potential, reduced cost and improved data reliability.
The freely available app solution was developed with farmers, technicians and coffee agronomists in Peru, Colombia and Honduras.
Croppie has scientifically demonstrated capacity to correctly count coffee cherries, works under field conditions, and can be used to identify predictors of coffee performance. Recent development of a smart data pipeline simultaneously improved the user experience and accuracy.
Contact us and explore our profiles
Romain Gautron
Postdoctoral Fellow
Christian Bunn
Scientist IIDownloads
Publications
Funders
Croppie was developed in collaboration with Producers Direct and Tecnicafé.
Croppie was a winner of the CGIAR Inspire Challenge 2020.
We are grateful for funding from the GIZ Fund for the Promotion of Innovation in Agriculture (i4Ag), GIZ FairForward – Artificial Intelligence for Development Cooperation, and the CGIAR Excellence in Agronomy and Sustainable Farming Science Program.