No traceability
When was that animal vaccinated? Nobody knows for certain.
Projects
Live
TerraCore PWA
TerraCore centralizes on a single screen everything a mid-sized farm needs to manage: livestock, inventory, production, animal health, costs, and finances. It replaces Excel, notebooks, and WhatsApp as operational tools. Built from the field with real producers in Urabá, it covers cattle, swine, equine, ovine, caprine, poultry, and associated crops (plantain, cacao, corn) in Colombia.
By Valentina Ramírez · Updated: June 19, 2026
Summary
Multitenancy PWA for mid-sized farms in Colombia. Centralizes livestock, inventory, production, health, and costs in one offline-first platform; cut administrative time by 42% in pilot farms. Django, DRF, PostgreSQL, React.
Stack
DjangoDRFPostgreSQLJWTReactTypeScriptTailwind CSS
The problem
Agricultural producers in Colombia had no software designed for them: existing ERPs were too complex, and an Excel sheet with login was not enough to manage livestock, animal health, and production simultaneously. Operations were scattered across notebooks, spreadsheets, and WhatsApp groups.
Invisible stock
The feed runs out and nobody saw it coming.
No signal, no data
There is no internet in the paddock. The record is left for later, and later never comes.
What I built
- 01
Multitenancy architecture: each organization operates in its own isolated data space on a single backend instance.
- 02
REST API with Django REST Framework, JWT authentication, and role-based access control across three roles: Administrator, Operator, and Collaborator.
- 03
Installable PWA from the browser on Android and iOS without going through the App Store: Service Worker persists records without signal and syncs in batch when connectivity is restored.
- 04
Six integrated modules: Dashboard with real-time KPIs, livestock (ID, breed, weight, health, location), supplies with minimum stock alerts, operating costs per batch, batch production, and animal health with vaccines and health history.
- 05
CSV import for frictionless onboarding and CSV export at any time: full data portability for the producer, no lock-in.
- 06
Relational modeling in PostgreSQL: organizations, users, animals, batches, supplies, and production records.
- 07
Frontend in React + TypeScript consuming the own API; responsive design with Tailwind CSS.
- 08
TLS on all endpoints; Ley 1581 Habeas Data compliance with no data mining for third parties and no model training with user data.
Dashboard
Real-time KPIs: active animals, monthly production, supplies, alerts, and estimated revenue.
Livestock
Complete record: ID, breed, weight, health status, and location by paddock.
Supplies
Feed, medications, and supplies with minimum stock alerts and supplier info.
Costs
Operating cost records per batch: supplies, labor, and expenses associated with production.
Production
Milk, egg, and crop batches with supply and hours traceability.
Animal Health
Vaccines, booster alerts, and health history per animal.
If you log a vaccine, three things happen automatically.
- Vaccine applied (e.g. Triple Bovina on 'La Canela')
- Supply automatically deducted (-1 dose from stock)
- Next booster scheduled (in 14 days)
- Animal health status updated
Architecture decisions
1. Offline-first with PWA
- Context
- In paddocks and pens the signal is intermittent or nonexistent.
- Trade-off
- A traditional web app would fail. A native app would require store publishing and more maintenance.
- Decision
- PWA with Service Worker that persists records locally and syncs in batch when connectivity returns. Installs from the browser on Android and iOS without going through the App Store.
2. Coupled modules with internal events
- Context
- Every field action has consequences on other records (vaccine, supply, health, schedule).
- Trade-off
- Isolated modules are easier to maintain but force the user to update each section manually.
- Decision
- Internal event architecture: one action triggers cascade updates across all related modules.
3. Multi-role without complexity for the user
- Context
- A farm has administrators, field operators, and collaborators with different information needs.
- Trade-off
- A single role sees everything (exposure risk and overload); too many roles complicate management.
- Decision
- Three roles with differentiated permissions: Administrator, Operator, and Collaborator. Each user sees only what they need based on their workflow.
4. Ley 1581 de 2012 compliance by design
- Context
- Production and animal health data are sensitive to the producer. Trust is a blocker.
- Trade-off
- Storing more data improves the product but increases legal liability and distrust.
- Decision
- No data mining for third parties, no model training with user data. TLS on all endpoints.
5. Guaranteed data portability
- Context
- Producers distrust platforms that retain their data or make exit difficult.
- Trade-off
- Data lock-in can retain users in the short term but destroys trust and slows initial adoption.
- Decision
- CSV import and export at any time with no restrictions. The producer's data belongs to the producer.
Results
-
-42% reduction in administrative task time measured in pilot farms.
-
Complete agricultural operation centralized on one platform, replacing the workflow in Excel, notebooks, and WhatsApp.
Learnings
-
Multitenant data isolation conditions every modeling decision: defining the tenant boundary early avoids rewrites.
One click away
Your next idea deserves
code that can carry it.
I design and build complete products: from the backend to the interface your users love. With integrated AI and security by design.
Projects from COP 2,000,000 / USD 500 depending on scope (MVP from 3-6 weeks).
Limited availability, I respond within 24h.
Start my project