StickyBot — The WhatsApp Sticker Revolution
Published on Aug 5, 2021
The Problem
When WhatsApp first introduced stickers, creating them was slow and complicated — users had to manually import images into third-party apps and package them into sticker packs.
A few early bots emerged to automate the process, but they quickly crashed under load once they hit a few thousand users.
The Idea
I set out to build a sticker bot that was not only easy to use, but also infinitely scalable.
The result was StickyBot — a WhatsApp bot that turned any image, GIF, or short video into a sticker in seconds.
Architecture & Development
StickyBot ran on a Kubernetes-based microservices architecture designed for reliability and speed.
System overview:
[Dummy Phone w/ WhatsApp]
↓
[TypeScript Service w/ WhatsApp Web Interface]
↓
[Core Microservice — Message Routing, User Logic]
↓
[MongoDB Queue]
↓
[Image & Video Processor — Sticker Generation]This setup allowed me to:
- Handle high concurrency using MongoDB as a lightweight message queue.
- Scale each component independently with Kubernetes HPA (Horizontal Pod Autoscaler).
- Automatically recover from node failures or overloads.
- Maintain low response latency (<1s avg) even under peak traffic.
Impact
StickyBot became a viral hit, reaching:
- 🧑🤝🧑 50,000 active users in just 2 weeks
- 🎨 600,000+ stickers created
- ⚡ 100% uptime, while competitors crashed after ~1,500 users
Eventually, Meta shut down the project due to WhatsApp’s anti-bot policy — but its stickers still circulate today, years after the service was turned off.
Even now, people occasionally send me stickers that were originally generated by StickyBot.