Know your leaves.
Protect your harvest.
Screen a coffee leaf with AI and optionally request an agronomist assessment.
How it works
3 STEPSChoose location
Choose department and municipality.
Check your photo
The AI flags visual patterns that may need attention.
Follow human review
Track a requested expert assessment here.
My requested assessments
ON THIS DEVICEAcademic prototype, not a Colombian government service. AI detections do not constitute a diagnosis.
Check a coffee leaf
Choose where the photo was taken before you scan it.
Add your leaf photo
JPG, PNG or WEBP · Maximum 8 MB
Submitting a photo stores a stripped copy, the AI result, municipality and date for historical demonstration reporting. Check our data notice before you continue. Do not upload people or sensitive data.
Learn about leaf damage
Three separate conditions supported by the detection model. Only roya is fungal; coco and minador are insect damage.
Roya · Coffee rust
Fungal leaf rust caused by Hemileia vastatrix. Check symptoms and institutional management guidance.
Coco · Weevil defoliation
Chewing damage associated with Curculionidae (Compsus / Epicaerus). Not a fungal infection. Seek a technician's identification before intervention.
Minador · Leaf miner
Leaf mining by Leucoptera coffeella, an insect pest. Learn to distinguish lesions from fungal spots and understand biological management.
Links lead to third-party institutional publications. The app does not prescribe pesticides; ask a qualified agronomist for local advice. Some historical publications may contain superseded chemical recommendations.
Authorities & verification
Public demonstration section for viewing uploaded demo images, recording a provisional assessment and downloading two separate historical reports.
Academic demonstration only. Public presentation mode hides individual photos and disables manual editing. Do not upload identifying information.
AI-generated findings and manual assessments are reported separately. Because images are submitted voluntarily rather than collected through representative sampling, these data cannot be used to estimate disease incidence.
Colombia reports
All uploaded images · AI candidates
Includes every accepted scan, whether reviewed or not.
Manually reviewed images only
Includes only cases assessed by the demo agronomist.
Reported departments
Shaded departments have at least one positive reported observation, not a proven outbreak. 32 departments and Bogotá D.C.
Cases for agronomist assessment
Built for Colombian coffee cooperatives
University research prototype exploring AI-assisted screening, separate manual reviews and responsible longitudinal reporting.
Three connected services
AI preliminary screening
Links to established agronomic research
Municipal reporting and human assessment
Data use & limitations
Accepted images are stripped of EXIF and stored with submission date and municipality. Default retention is 90 days for photos and 365 days for records in this academic prototype. Request deletion from your Home card or an administrator. Processing requires notice acceptance and an appropriate legal basis in production.
Feedback & Bug Reports
Help improve CoffeeLeaf. Share an observation or report a problem you encountered.
Not an official or accredited Colombian public service. Reviewers are demonstrators until their professional credentials are independently confirmed.