Inspiration We saw thousands of Pakistani "Instagram Sellers" struggling to manually reply to repetitive messages like "Price?" and sharing IBANs. Existing tools were either too expensive (Official API) or too complex. We wanted to build a bridge: enterprise-level AI automation for the casual seller, at zero cost, speaking their local language (Roman Urdu).
What it does WaShop.online turns a standard WhatsApp number into an automated store:
AI Sales Agent: Instantly replies to queries in Roman Urdu/English and guides users to checkout. Status Automation: Automatically posts new products to WhatsApp Status for organic marketing. Unified Dashboard: Manages orders, sales, and analytics in one clean interface. Local Payments: Built-in verification for Easypaisa, JazzCash, and bank transfers. How we built it We used the MERN Stack (MongoDB, Express, React, Node) for performance:
Backend: Node.js with whatsapp-web.js for a "headless" WhatsApp client (bypassing paid API costs). AI: integrated Google Gemini Pro with custom prompt engineering for Roman Urdu sales. Frontend: React + Vite with Tailwind CSS and Shadcn UI for a premium aesthetic. Real-time: Socket.io for instant syncing between WhatsApp and the web dashboard. Challenges we ran into Session Stability: Keeping the headless WhatsApp session alive in Docker was tricky and required robust auto-reconnect logic. Language Nuance: training the AI to distinguish between casual slang and serious inquiries in Roman Urdu (e.g., "bhai kam kardo"). Docker Networking: Getting Nginx, Backend, and WebSockets to talk correctly inside containers took several iterations. Accomplishments that we're proud of Zero-Cost Capable: Bringing automation to the masses without requiring expensive Meta API keys. The UX: creating a dashboard that looks like a high-end SaaS product, not a prototype. Local Fluency: Seeing the bot successfully haggle and close deals in Roman Urdu. What we learned Prompt Engineering: Writing effective system prompts is just as critical as writing code. State Management: Handling complex conversation flows requires careful logical planning. Open Source: The incredible value of community libraries like whatsapp-web.js. What's next for WaShop.online Voice Notes: Adding Whisper AI to understand and reply to audio messages. Dual Mode: Offering an upgrade path to the Official WhatsApp API for larger businesses. Inventory Sync: One-click syncing with Facebook Marketplace and Shopify.
Built With
- css3
- docker
- express.js
- git
- google-gemini
- html5
- javascript
- mongodb
- nginx
- node.js
- openai
- react
- restapi
- shadcn-ui
- socket.io
- tailwind-css
- typescript
- vite
- whatsapp-web.js
Log in or sign up for Devpost to join the conversation.