About

We build the farming information that was missing in Somali.

Beeroole is an agricultural technology company based in Hargeisa, Somaliland. We make a mobile app and a public library of crop guidance for smallholder farmers across Somaliland and Somalia.

What we make

The Beeroole Android app does four things: it identifies crop diseases from a photograph, publishes a written guide for each disease it knows about, lists market prices collected from local markets, and delivers early warnings issued for a farmer's region.

Everything in the app's disease library is also published on this website, free to read and free to link to. We publish it openly because an agronomist, a teacher or an extension officer should be able to send a farmer a link without asking them to install anything first.

Who it is for

Smallholder farmers growing vegetables and cereals — tomato, onion, pepper, cabbage, spinach, cucumber, potato, maize and beans — on plots where a single disease outbreak is the difference between a season's income and none.

It is also for the people who advise them: agricultural offices, cooperatives, NGOs and development programmes who need reference material in the language the farmer actually speaks.

How the disease guides are written

Each guide is compiled from published agricultural extension material — university extension services, plant pathology references and recognised horticultural sources — and then adapted to the crops, climate and inputs that are realistically available here.

Every guide records, on the page:

  • The sources it was written from, linked, so you can check them.
  • How confident we are in it — from "checked against published extension sources" down to "uncertain, treat as a starting point".
  • Whether an agronomist has reviewed it, stated plainly either way.
  • The date it was last updated.

Where a disease is genuinely ambiguous — where two pathogens produce the same spots on the same leaf — the guide says so and links to the ones it is confused with, rather than picking one and sounding certain.

How we use image recognition

The app compares a photograph against the diseases in its library and suggests the closest match with a confidence score. That is a genuinely useful narrowing step. It is not a diagnosis, and we do not present it as one.

Photographs are a limited signal: several diseases look alike, light and camera quality change what the model sees, nutrient problems mimic pathogens, and a plant can carry more than one problem at once. So the app puts the written guide beside the suggestion, shows the diseases it is easily confused with, and tells the farmer when the answer is to speak to a person instead.

Corrections

If you are an agronomist, plant pathologist or extension officer and you find something wrong in a guide, please tell us — email info@beeroole.com with the page and what is incorrect. Corrections to agricultural advice take priority over everything else we are working on.