Building CropChain OS: An AI-Powered Financial Operating System for India's FPOs with AWS Aurora DSQL & Vercel
Discover how we designed and built a cloud-native platform that helps Farmer Producer Organizations (FPOs) optimize mandi decisions, automate post-harvest operations, and deliver transparent farmer settlements #H0Hackathon
Meet the Team
CropChain OS was built by a team of two during the H0 Hackathon.
I focused on the backend architecture, AWS Aurora DSQL integration, relational database design, APIs, and cloud infrastructure, while my teammate @karthikam14 led the product vision, UI/UX, frontend development, user workflows, dashboards, and overall application experience.
Building a complete production-style platform in just a few days meant constantly collaborating on architecture, APIs, database modeling, AI integration, testing, and user experience. Every major feature was shaped through discussions and rapid iteration, making this one of our most rewarding hackathon projects.
Where the Idea Came From
When we started researching India's agricultural ecosystem, one thing became obvious.
Technology has helped farmers improve crop production, but the post-harvest journey remains largely manual.
India has over 35,000 Farmer Producer Organizations (FPOs), including 10,000+ government-supported FPOs, representing more than 30 lakh farmers.
Yet many FPOs still depend on:
- WhatsApp groups
- Excel spreadsheets
- Paper records
- Manual bookkeeping
- Phone calls
This creates major operational inefficiencies.
FPO managers struggle to decide where to sell crops for the highest real return, while farmers often have no visibility into where their harvest was sold, how settlements were calculated, or when payments will arrive.
We kept asking ourselves one simple question:
What if every FPO had its own operating system?
That question became CropChain OS.
What We Built
CropChain OS is an AI-powered Financial Operating System designed specifically for Farmer Producer Organizations (FPOs).
Instead of solving one isolated problem, we digitized the complete post-harvest lifecycle.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Farmer Registration
↓
Harvest Submission
↓
Godown Verification
↓
AI Mandi Recommendation
↓
Dispatch Tracking
↓
Sale Recording
↓
Automatic Farmer Settlement
↓
Financial Ledger
↓
Analytics & Credit Score
Every harvest can now be tracked from submission to final farmer payment.
CropChain OS helps FPO managers optimize operations while giving farmers complete transparency throughout the selling process.
Designing Two Different Experiences
One of our biggest design challenges was building for two completely different users.
FPO Manager Dashboard
Managers need operational visibility.
The dashboard provides:
- Harvest Management
- AI Mandi Optimizer
- Live Mandi Map
- Dispatch Tracking
- Sales Recording
- Automated Payouts
- Revenue Analytics
- Credit Score Dashboard
Farmer Portal
Farmers need simplicity.
The mobile-first portal allows farmers to:
- Join an FPO
- Submit harvests
- Track dispatches
- View settlement breakdowns
- Monitor payment history
- Access multilingual support
Everything was designed to require minimal technical knowledge while providing complete transparency.
Building on AWS Aurora DSQL
One of the most important decisions we made was choosing AWS Aurora DSQL as the foundation of CropChain OS.
Rather than storing isolated application data, we designed a relational database that models the real operational workflow of an FPO.
Our database consists of 17 interconnected tables, including:
- Managers
- FPOs
- Farmers
- Memberships
- Harvests
- Mandis
- Mandi Prices
- Trust Scores
- Dispatches
- Sales
- Payouts
- Events
- Ledger
- Credit Scores
- Analytics
- Notifications
This architecture ensures every crop movement remains completely traceable—from harvest submission to the final farmer settlement.
Working with Real Government Data
We wanted CropChain OS to solve real problems, so we avoided mock datasets.
Instead, we imported official AGMARKNET agricultural market data into Aurora DSQL.
Our platform currently includes:
- 1,614 real mandi locations
- 3,657 official government price records
Using authentic government datasets allowed us to build realistic AI recommendations and simulate actual agricultural workflows.
AI That Explains—Not Decides
One design principle guided our AI implementation:
Financial decisions should remain deterministic.
Instead of asking an LLM to calculate prices or payouts, CropChain first computes:
- Transport costs
- Commission charges
- Payment reliability
- Trust scores
- Expected net returns
The highest-ranked mandis are then passed to Google Gemini, which explains why a recommendation was made in natural language.
This approach makes every recommendation:
- Explainable
- Reproducible
- Auditable
- Easy for users to trust
Technology Stack
Frontend
- Next.js 15
- React 19
- TypeScript
- Tailwind CSS
- Framer Motion
Backend
- Next.js API Routes
- Node.js
Database
- AWS Aurora DSQL
AI
- Google Gemini 2.5 Flash
Maps
- Leaflet
- React Leaflet
Visualization
- Recharts
Deployment
- Vercel
Challenges We Faced
Every ambitious hackathon project comes with challenges.
Some of our biggest were:
- Designing a relational schema that accurately models real FPO workflows.
- Integrating AWS Aurora DSQL while handling distributed SQL behavior.
- Importing thousands of real government market records.
- Building two different user experiences without duplicating backend logic.
- Combining deterministic business rules with AI explainability.
- Delivering a polished, production-style application within the hackathon timeline.
Each challenge ultimately strengthened the platform.
What We Learned
CropChain OS reinforced something we've always believed:
Great software starts with understanding real-world problems.
We learned that thoughtful database design shapes the entire application, AI is most valuable when paired with deterministic systems, and successful full-stack products require frontend, backend, cloud infrastructure, and user experience to evolve together.
Perhaps our biggest takeaway was that solving meaningful problems isn't just about writing code—it's about building trust.
What's Next
CropChain OS was built during a hackathon, but it was designed with real deployment in mind.
Our roadmap includes:
- Live AGMARKNET price synchronization
- Automated UPI farmer payouts
- Voice-based multilingual assistant
- AI price forecasting
- WhatsApp harvest submissions
- Multi-FPO benchmarking
- Credit & lending integrations
- Predictive operational analytics
Final Thoughts
India has already invested heavily in strengthening Farmer Producer Organizations.
We believe the next step is providing them with software that is just as modern as the agricultural ecosystem they support.
CropChain OS aims to become that operational layer—helping FPOs make smarter selling decisions while giving every farmer complete confidence in every harvest, every sale, and every payment.
This is my first technical article, and I hope it's the first of many.
I'm excited to keep building, keep learning, and share more projects, technical write-ups, and lessons from my journey.
Thanks for reading, and see you in the next build! 🚀
Built for the H0 Hackathon.
#H0Hackathon
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article