Inspiration:

Marketing teams drown in fragmented data — Google Ads in one tab, Meta in another, email in a spreadsheet, organic in yet another dashboard. I wanted to build a tool that kills that pain : one unified view and one AI YOU can just ask.

What it does :

Nexlytic pulls every marketing channel into one single normalized schema powered by amazon aurora dsql . real time dashboard shows the revenue , spend , ROAS and conversions side by side. And an AI layer let's you ask questions in plain english and get answers grounded in your live data and not generic guesses.

How I built it :

I used next.js with the pages router for both the font end and API layer, the database is amazon aurora dsql (serverless PostgreSQL-compatible) with IAM token-based authentication — no stored passwords. Styling is fully custom CSS with a dark theme and lime green accent system. and it was all built within 10days with a lot of trial and error

Challenges I ran into:

Getting the design and styling to match my vision was a constant battle ,I iterated through multiple approaches to achieve the dark premium aesthetic with the right balance of contrast, spacing, and micro-interactions.

Accomplishments that I'm proud of :

I was able to build a fully functional demo before the deadline — solo, from zero to a working product with live database, AI integration, and polished UI. I'm also proud that the AI feature actually queries real data rather than returning canned responses , it reasons over live Aurora DSQL rows, which means every answer is grounded in truth.

What I learned :

I learned how to wire up Aurora DSQL with IAM-based auth tokens in a serverless context something with very little documentation out there. I also deepened my understanding of building premium UI without relying on component libraries, EVERY animation , hover state , transition was handcrafted and i am proud of it

What's next for Nexlytic

Multi-tenant onboarding so teams can connect their own ad accounts via OAuth. Real-time anomaly alerts pushed to Slack/email when a channel's metrics deviate. A multi-touch attribution engine that shows the true customer journey across channels. And a mobile companion app so marketing leads can check performance on the go.

Built With

  • amazon-aurora-dsql
  • boxicons
  • claude
  • css
  • next.js
  • vercel
Share this project:

Updates