Context
ORCEI is a private project for generating professional quotes that can be shared through WhatsApp. I am keeping the repository private, but the product decisions are useful as a portfolio case because the system sits in the messy overlap between mobile usage, client communication, and document-style outputs.
Problem
Small service businesses often still create quotes manually or in ad hoc documents. The product needed to make quote generation fast, mobile-friendly, and easy to share with clients through a public link and WhatsApp.
Constraints
- The app needed to feel comfortable on mobile first.
- Public quote views had to work without exposing private account data.
- AI assistance had to stay bounded to one narrow workflow: improving service descriptions.
- The team surface area was small, so the system needed a compact stack.
Architecture
authenticated app
-> quote wizard
-> Supabase-backed profiles, clients, quotes, quote_items
-> public /q/[token] route for client-facing quote views
-> WhatsApp sharing utilities
Decisions and trade-offs
- I used Supabase with RLS because the project benefits from quick iteration and clear per-user data boundaries.
- The AI feature is intentionally narrow; it improves quote descriptions instead of trying to run the whole workflow.
- Printing support shipped before server-side PDF generation because that solved the user problem faster.
What worked
- The route structure makes the product legible: authenticated app, auth area, public quote pages, and small server actions.
- The feature checklist shows a healthy product shape already: profile, settings, public links, dashboard, and share actions.
- The documentation around environment setup and database schema made it easy to keep the project maintainable despite staying private.
What is still incomplete
- Editing, duplication, and deletion flows for quotes were still planned at the time of the last update.
- Offline support and server-generated PDF downloads were not complete.
- This should eventually evolve into the stronger public-facing concept I now call
QuoteFlow BR.
Evidence
Implemented:
- 3-step quote wizard
- public token route for quote sharing
- WhatsApp share utilities
- AI endpoint for improving descriptions
- mobile-first layout and PWA shell
Database core:
profiles -> clients -> quotes -> quote_items