Project Case Study · 2025
An AI-powered bilingual platform built for Indian farmers — delivering crop intelligence, disease detection, mandi prices, and government scheme discovery in one place.
Matruka is an AI-powered smart farming platform built specifically for Indian farmers, bridging the gap between cutting-edge agricultural technology and rural accessibility.
The platform operates entirely in both English and Hindi, ensuring farmers who aren't comfortable in English can still access every feature. From real-time weather and crop recommendations to uploading soil profiles and detecting plant diseases through images — Matruka puts an agronomist in every farmer's pocket.
Beyond AI features, Matruka aggregates real mandi (market) prices, locates the nearest mandis, and surfaces relevant government agricultural schemes — data that is often fragmented across dozens of portals and completely inaccessible to most farmers.
Built as a solo full-stack project, every layer — UI design, backend architecture, AI integrations, API aggregation, and deployment — was designed and shipped by a single developer.
Enter your location and the platform fetches live weather, cross-references soil conditions, and recommends the optimal crops for the current season using AI models.
AI · Open Weather API · Soil DataFarmers can photograph or upload their soil test documents. The system parses and analyzes the profile to suggest suitable crops and fertilizer strategies.
Image Processing · AI AnalysisUpload a photo of a diseased crop leaf. The AI model identifies the disease, confidence level, and recommended treatment — all within seconds.
Computer Vision · ML ModelA curated, regularly updated database of Central and State government agricultural schemes — surfaced based on the farmer's state and crop type.
Data Aggregation · Official Govt SitesReal-time commodity prices from mandis across India, aggregated from government data sources. See price trends and find the best market for your produce.
Live Data · Commodity OnlineGeolocation-powered nearest mandi finder — shows distance, current prices, and operating hours so farmers can plan their market visits efficiently.
Geolocation · Google MapDesigned for low-literacy users in rural India — large touch targets, bilingual labels, high-contrast imagery, and a clean card-based layout that works on budget Android phones.
AI Crop Recommendations
Soil Profile Upload
Disease Detection
Government Schemes
Live Mandi Prices
Mandi Near Me
Clean clutter free UI - minimal and clean. Supirior performance and creative freedom.
Express server handling API proxying, AI calls, and data aggregation from external sources.
Flexible document storage for schemes, mandi records, and user session data.
Gemini 2.5 Flash vision models for disease detection; language models for crop advisory generation.
Real-time meteorological data powering location-aware crop recommendations.
A site which shows live crop prices of your city, town or village.
Modern async/await patterns, optional chaining, and module-based architecture throughout.
Deployed with zero-config CI/CD — automatic preview deployments on every push.
No personalized crop advice without an agronomist Farmers in rural India can't afford or access agricultural experts.
Matruka uses Gemini 2.5 Flash to analyze their location's real-time weather and give crop-specific recommendations instantly — no expert needed.
Government schemes go completely unclaimed Billions in agricultural subsidies go unused every year because farmers don't know they exist.
Matruka surfaces official government scheme links directly — no middleman, no confusion, no missed benefits.
Mandi prices are a guessing game. Farmers often sell at a loss because they don't know what the same crop is fetching at a market 30km away.
Commodity Online's real-time price feed shows live rates across markets so farmers can negotiate from a position of knowledge.
Agri-tech tools are English-only. Most farming apps assume English literacy, cutting out the majority of India's farming population.
Matruka is fully bilingual — every feature, every label works in Hindi.
The live site is deployed on Vercel. The source code is on GitHub.
Both links are right below.