AgriVerse

Helping farmers grow wealth, not just crops.

Comment

Inspiration

The core issue in Indian agriculture is the lack of actionable intelligence to counter fragmented markets, unpredictable weather, and crop diseases. Existing digital tools offer only static data. Our inspiration was to build a single, intuitive platform (AGRIVERSE AI) that moves beyond raw data to deliver Prescriptive, Localized AI Advice, thereby helping smallholder farmers maximize net profit and build resilience.

What it does

1)AI Advisor (Assistant): Provides immediate diagnosis (image upload) and optimizes selling decisions (when/where to sell based on net profit margin).

2)Market: Offers real-time price monitoring and integrated Profit & Loss (PNL) analysis.

3)Planning: Delivers proactive schedules for crop management, including Smart Irrigation and Soil/Fertilizer Recommendations.

4)Finance: Features a personalized Loan & Subsidy Hub (with repayment alerts) and an interactive Cash Flow Tracker (with visual expense breakdown and receipt upload).

5)Schemes: Provides a filterable database of government grants and schemes based on the farmer's land size and eligibility.

How we built it

We used React Native for cross-platform stability (Web/Mobile) and a clean, modular architecture. Core Structure: Built on a single-state system with Tabbed Navigation for easy scaling. AI Engine (Simulated): Developed custom dynamic algorithms to simulate real-time LLM advice for selling (geospatial/profit) and financial guidance. Accessibility: Ensured full 5-language localization for Hindi, Telugu, Tamil, Malayalam, and English.

Challenges we ran into

  1. Data Granularity and Integrity (Real-World Data Problem) The primary challenge was simulating the granularity and real-time integrity required for prescriptive AI. In reality, data needed for core features—like Soil Health, Pest Incidence, and Mandi Prices—varies drastically by village, not just district.

Impact: Our system had to rely on generalized mock data (e.g., using one market price for an entire region), which limits the immediate precision of the AI Selling Advisor's recommendations. This highlights the future need for robust IoT (Internet of Things) integration to capture hyper-local farm data.

  1. Geospatial Visualization vs. Stability (Technical Trade-off) Achieving a true visual map interface (for the AI Selling Advisor's location/distance feature) proved impossible within the constraints of a lightweight cross-platform environment.

Solution Trade-off: We were forced to pivot from an interactive map module to a stable Geo-Visualizer (simulated via structured text and CSS). This trade-off prioritized app stability and instant loading times over complex external map library dependencies.

  1. Localization of Technical and Financial Terms While multi-language support (English, Hindi, Telugu, Tamil, Malayalam) was achieved, ensuring the accurate translation of specialized terminology was complex:

Terminology Issue: Concepts like "Net Profit Margin," "PM-KISAN," or specific "DAP Fertilizer" were handled via direct key replacement. This requires constant human verification to ensure the local agronomic and financial jargon is culturally and technically correct for the farmer in each region.

Accomplishments that we're proud of

Full Localization: Delivering a truly functional, five-language interface that makes advanced technology accessible to millions of non-English speaking farmers.

Actionable Intelligence: Successfully integrating complex, high-value AI features (AI Selling Advisor and Soil/Irrigation Management) that generate immediate, prescriptive advice.

Unified Financial Hub: Creating a centralized Finance tab that combines loan tracking, repayment alerts, cash flow logging (with mock OCR integration), and PNL analysis in one place.

What we learned

We learned that in agricultural technology, the best UX prioritizes speed, clarity, and trust over complexity. Simulating real-time intelligence using highly realistic mock data (like the dynamic Mandi recommendations) can be a powerful proof-of-concept when live API integrations are not feasible, allowing immediate validation of the value proposition.

What's next for AgriVerse

Gemini/Bedrock API Integration: Replacing all current mock AI data with real-time calls to the Amazon Bedrock AI Agent for live, geographically accurate price predictions, pest diagnosis, and advanced financial modeling.

Hardware Integration: Developing modules to integrate with Soil IoT Sensors to pull live NPK/pH data directly into the Soil & Nutrition Management section.

E-Mandi Integration: Establishing partnerships with local e-Mandi platforms to allow farmers to sell their recommended crop directly through the app interface.

Built With

Share this project:

Updates