Full Stack · AgriTech
A smart-agriculture platform built for Kenyan farmers: farm and crop records, weather and drought alerts, a produce marketplace with real market prices, M-Pesa payments, and expert advisory content — with an offline mode for the many places where connectivity can't be assumed.
Demo status: not currently hosted, but the repository ships a Dockerfile and detailed installation and feature documentation, so it stands up locally in a few commands. Setup steps are at the bottom of this page.
A broad platform, deliberately — the value to a farmer comes from the pieces working together.
A smallholder farmer in Kenya makes decisions with incomplete information: what to plant and when, whether rain is coming, what maize is fetching at the nearest market this week, and whether a spot on a leaf is worth worrying about. Each answer lives somewhere different, and most of those places assume a good internet connection.
Kilimo Guru puts the farm record, the forecast, the market price and the advice in one platform, and keeps working when the connection drops.
Sole developer. I designed the domain model across farmer registry, crops, marketplace, advisory and analytics, built the Django apps and REST layer, and containerised the project with documentation thorough enough for someone else to run it.
Each app owns one part of a farmer's working life.
Farmer registration, profiles and GPS-mapped farm records.
Planting calendars by region, yield tracking and rotation planning.
Expert articles, consultations and e-learning content.
Yield forecasts, profitability reports and market trends.
REST serializers and endpoints for the mobile and offline clients.
Produce listings, buyer matching and real-time market prices.
Forecasts, drought and flood warnings, irrigation advice.
M-Pesa payments, farm credit scoring and crop insurance.
One-command environment so setup isn't the barrier to trying it.
The target user regularly has no signal. Building as a PWA with local storage and sync-on-reconnect means the app degrades to slower, not broken.
Card payments would exclude most of the intended users. Mobile money is how this market actually pays, so it's the primary integration rather than an alternative.
The repo carries installation, feature, structure and completion documents. On a platform this broad, an undocumented feature is functionally a missing one.
Advisory content supports Swahili and local languages at the data layer, so translation isn't a UI afterthought.
Serializers and endpoints live in their own app rather than scattered across feature apps, which keeps one clear contract for every client.
A Dockerfile from the start means the same environment locally and on any host, which matters most on a project with this many moving integrations.
The fastest way to see it working end to end.
git clone https://github.com/Eunice-ctrlz/KILIMOGURU.git
cd KILIMOGURU
python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
python manage.py migrate
python manage.py runserverRuns on localhost:8000. The repo's installation guide covers the optional integrations.
I'm happy to talk through the architecture or the decisions above — and I build systems like this for other people too.