Inspiration

Navigating complex prescription schedules is a massive challenge, especially for patients dealing with multi-dose treatments or chronic illnesses. Mismanaged medications lead to thousands of complications daily. I built Remba to bridge this gap, transforming the complexity of medical prescriptions into clear, personalized, and interactive WhatsApp schedules that keep patients safe, informed, and on track.

What it does

Remba acts as an AI-powered medication companion directly within WhatsApp. A user simply types out their doctor's instructions in plain language. Remba's extraction backend parses the medical data, automatically schedules the proper intervals, tracks user history, and initiates precise check-ins to make sure no dose is missed.

How we built it

The platform is engineered with a high-performance, production-grade architecture:

  • Backend: Built using NestJS with a modular, highly scalable design pattern.
  • Database & ORM: Powered by PostgreSQL, managed seamlessly with Prisma ORM for real-time state persistence and rapid transaction modeling.
  • Integrations: Connected directly via the official Meta WhatsApp Business Cloud API using custom routing mechanisms, webhooks, and automatic session-state synchronization.
  • Hosting: Fully containerized and deployed in production on Railway for fast, unbuffered webhook request delivery.

Challenges we faced

Transitioning from a local environment to the cloud threw a massive curveball. I spent over 6 days fighting deep cache locks inside Meta's background event-routing architecture, where synthetic testing payloads from Postman succeeded but real-world mobile provider packets were dropped. Overcoming this required engineering a manual Graph API override protocol to forcefully bind the WhatsApp Business Account metadata layers directly to our cloud destination gateways. Another challenge is publishing the app and making it public, since I'm not a business, it's been hard to find a way to make every number that writes to Remba whitelisted

Accomplishments that we're proud of

I successfully bypassed every networking layer constraint to build an open, responsive, and functional end-to-end cloud pipeline. The moment I saw a physical text message bypass Meta's global infrastructure firewalls and map flawlessly onto my NestJS routing controllers in real-time was a massive engineering victory.

What we learned

I gained an immense, practical understanding of asynchronous network flows, third-party webhook handshakes, and database persistence layers under strict validation criteria. Most importantly, I learned that real-world deployment requires deep resilience and a methodical approach to infrastructure debugging.

Note

Because Remba runs in Meta's secure sandbox developer environment, Meta requires that your phone number be whitelisted before you can chat with the live bot. Click this link to open the chat thread, drop your number in our submission comments, and we will instantly whitelist you to test it live! Alternatively, please watch our attached video demo to see the Railway backend tracking medical queries in real time.

Built With

Share this project:

Updates