AgriProduct

Project Overview

AgriProduct is an online agriculture marketplace built around three different users: farmers who need supplies and advice, shopkeepers who want to sell agricultural products, and admins who keep the whole platform trustworthy.

I built it end to end with Next.js and MongoDB, combining multi-vendor e-commerce with AI-powered farming advice, real-time weather data, and a full admin approval system.

The Challenge

Most agriculture e-commerce sites solve one problem: letting someone buy a product online. AgriProduct needed to solve several at once.

Farmers needed more than a shopping cart. They needed a way to identify crop diseases and pests, get advice they could actually understand (in Urdu as well as English), and know their local weather before making decisions. Shopkeepers needed real store management tools, not just a product upload form. And because anyone can claim to be a seller, the platform needed a way to verify shops and handle complaints before a bad actor could damage trust in the whole marketplace.

The Solution

I built AgriProduct around three role-based dashboards instead of one generic account type. Farmers get shopping, AI crop analysis, weather, and a complaints system. Shopkeepers get sales analytics, inventory, and order management. Admins get platform-wide oversight: user management, shop approval, and complaint resolution.

The Marketplace Homepage

The homepage leads with a clear value proposition, "Buy Agri Products Online," backed by a "Why Choose AgriProducts?" section that sets expectations early: verified shops and farmers, AI-assisted decisions, a simple experience for non-technical users, and support in both Urdu and English.

Below the fold, real product listings (seeds, fertilizers, tools, even equipment like tractors) are shown with live pricing in PKR, alongside platform stats and seasonal farming tips for Spring through Winter.

AI-Powered Farming Advice

The Consultation page is where AgriProduct stops being a store and starts being an advisor. Three AI services are offered directly: Crop Disease Detection from an uploaded photo, Pest Detection & Control with eco-friendly recommendations, and Weather-Based Advice for irrigation and harvesting timing.

A simple three-step explanation (input your data, AI analyzes it, receive recommendations) makes the AI feel approachable instead of like a black box, which matters for a user base that may not be deeply technical.

An AI Assistant That Speaks Urdu, Too

Alongside photo-based crop analysis, a floating AI Chat Assistant is available anywhere in the app. It opens with common starter questions ("How to improve soil fertility?", "Best fertilizers for wheat?") and accepts typed or spoken input, with a language switcher so a farmer can ask in Urdu just as easily as English.

That language flexibility isn't a minor detail. For a lot of the platform's real users, it's the difference between a tool they can actually use and one they can't.

The Farmer Dashboard

Once logged in, a farmer's dashboard becomes their home base: total orders, a local weather alert, and open complaints are visible immediately, alongside a full local weather widget with wind, humidity, precipitation, and UV index.

Shortcuts to the most useful actions, crop analysis, weather, nearby shops, and filing a complaint, sit front and center, so a farmer never has to hunt through menus to get help.

Running A Shop

Shopkeepers get a genuinely useful business dashboard, not just a product list. Total sales, orders, customers, and product count are tracked alongside a sales trend chart, recent orders with delivery status, and an inventory table that flags stock and pricing at a glance.

Every order can be processed and every low-stock item updated directly from this view, so running a shop on AgriProduct doesn't require jumping between separate tools.

Listing A Product

Adding a new product uses a proper multi-step form, Basic Info, Pricing, Shipping, Images, Inventory, Variants, and Advanced settings, each as its own tab, with a rich text editor for the product description.

Breaking a complex form into tabs kept the shopkeeper experience from feeling overwhelming, while still supporting real e-commerce needs like variants and shipping rules.

Keeping The Marketplace Trustworthy

None of this works if the marketplace isn't trustworthy, so admins get their own dashboard: total users, active shops, open complaints, and total products, backed by charts for user growth, order distribution, and shop registrations over time.

New shops don't go live automatically. A Shop Management view lets admins filter by status (All, Approved, Pending, Rejected) and review each shop's owner, city, and contact details before approving it.

The same admin dashboard handles user complaints with severity levels (Low, High) and a status pipeline (Pending, Processing, Resolved, Rejected), so problems get triaged instead of piling up unanswered.

Content Structure

AgriProduct is organized around three roles sharing one platform.

The app includes clear areas for:

• Product browsing and checkout
• AI crop analysis and chat assistant
• Weather and seasonal advice
• Farmer orders and complaints
• Shopkeeper store management
• Admin user, shop, and complaint oversight

Keeping each role's tools cleanly separated made it possible to add features for one role (like the AI assistant) without cluttering the experience for the others.

The Result

AgriProduct brings together e-commerce, AI-powered farming advice, and platform trust and safety into a single application. Farmers get real answers to real problems in the language they're comfortable with, shopkeepers get proper business tools instead of a bare product list, and admins have the oversight needed to keep the marketplace trustworthy as it grows.

Key Takeaways

The hardest part of this project wasn't any single feature, it was designing three genuinely different experiences (farmer, shopkeeper, admin) on top of one shared data model without any of them feeling like an afterthought.

For AgriProduct, that meant treating role-based access as a first-class part of the architecture from the start, and making sure features like AI crop analysis and Urdu voice support actually matched how the platform's real users would use it, not just how a typical e-commerce demo would.