Building Grahbook: WhatsApp Commerce for 600M+ Indian Merchants
How I built a full-stack WhatsApp commerce platform that turns any WhatsApp number into an AI-powered store with UPI payments, order management, and automated ledger entries.
The Problem: 600M+ Merchants Trapped in WhatsApp
India has over 600 million small businesses — kiranas, street vendors, local shops — that run entirely on WhatsApp. They chat with customers, take orders, share photos, and close deals all inside WhatsApp. But there's a massive gap: **zero tools** to manage orders, payments, or inventory within the app they already use.
Merchants juggle spreadsheets, paper notebooks, and manual UPI screenshots. Revenue tracking is impossible. Order management is chaos.
Why I Built Grahbook
I wanted to solve a real problem for real people. Grahbook turns any WhatsApp number into an AI-powered store — customers can browse catalogs, place orders, and pay via UPI without ever leaving the chat.
The Tech Stack
- **React + TypeScript** for the merchant dashboard
- **Vite** for lightning-fast builds
- **Framer Motion** for buttery animations
- **Firebase** for real-time data sync
- **TailwindCSS** for the design system
Key Features
AI-Powered Ledger Every conversation is automatically parsed into structured ledger entries. Revenue tracking happens in real-time without manual input.
WhatsApp-Native Interface Customers don't need to download anything. They chat with the merchant's existing WhatsApp number and get a catalog experience.
UPI Payment Integration Payments happen through the same UPI flow merchants already use, but now with proper order linking and reconciliation.
Results
Grahbook has been deployed and is being actively iterated on. The merchant dashboard shows AI-powered ledger entries, automated revenue tracking, and a bento-grid metrics layout — all animated with Framer Motion.
Lessons Learned
- **Start with the user problem, not the technology.** I didn't pick React because it's trendy — I picked it because the merchant dashboard needed real-time updates.
- **AI should solve problems, not impress investors.** The ledger AI is invisible to merchants but saves them hours every week.
- **Deploy early, iterate fast.** The first version was rough. But real users gave better feedback than any spec document.