Jualoka - UMKM E-Commerce Platform
A lightweight e-commerce platform for Indonesian UMKM to create online stores and receive orders directly via WhatsApp. Includes AI-powered product analytics and a real-time order notification system.
Problem & Objective
Most Indonesian UMKM sellers rely on informal channels like WhatsApp broadcasts and Instagram DMs to take orders, leading to missed messages, stock confusion, and no visibility into which products actually sell well. Existing e-commerce platforms are too complex and charge fees that eat into already-thin margins.
Jualoka was built to give small business owners a store presence they can set up in minutes, with orders flowing directly into their WhatsApp without any new app for customers to download. The platform also analyzes sales data automatically so sellers know which products to produce more of and which to cut.
Challenges & Solutions
1. Real-Time Order Notifications Without a Third-Party Service
Sellers needed to know the moment an order arrived without relying on push notification services. I implemented Server-Sent Events (SSE) on the API layer so the admin dashboard maintains a live connection to the server and receives instant order alerts the moment a customer submits a cart, without polling or external dependencies.
2. WhatsApp Checkout Without an API or Integration
Integrating a formal WhatsApp Business API would require approval and add cost. Instead, I engineered a checkout flow that composes a fully formatted order message from the cart state and opens it directly in WhatsApp via the wa.me deep link. Customers need no app install, no login, and sellers receive orders in their existing WhatsApp number.
3. AI Product Performance Classification
Sellers needed actionable insight on their product catalog without understanding data. I built an analytics module that calculates revenue, volume, and profitability per product, then uses Gemini to classify each product as Laris, Kurang Laku, or Rugi with a plain-language recommendation on what to do next, making data accessible to non-technical business owners.
Results & Impact
- Sellers can create a fully functional online store with a unique public URL in under five minutes.
- The WhatsApp checkout flow eliminates friction for buyers while keeping the entire order process within tools sellers already use daily.
- Real-time SSE notifications mean sellers never miss an order even when the dashboard tab is open in the background.
- AI product classification gives UMKM owners data-driven production decisions that previously required a business analyst to produce.
Specific Role
Full Stack Developer
Built the complete platform including the public storefront, seller dashboard, product and order management system, SSE notification layer, WhatsApp checkout flow, Vercel Blob image storage integration, and Gemini-powered analytics module.
Tech Stack
Status
Completed / Active