All projects

Case study

Masar X

Bilingual AI study platform: semantic PDF search and vector document indexing across web, desktop, and mobile clients on a single Supabase backend.

Live demoSource code
  • Next.js
  • TypeScript
  • Supabase
  • pgvector
  • Electron
  • Expo

The problem

Students need to search and study from large PDF collections quickly, on whatever device they have, without their materials and progress living in separate silos.

What I built

A pnpm monorepo sharing one Supabase backend (RLS, Edge Functions) across a Next.js web app, an Electron desktop app, and an Expo mobile app, with semantic search over ingested PDFs and vector document indexing.

Last updated November 7, 2026