Pavly
SaaS · founder build · built by Aziz Khan
The problem
Small shops sell through reels and DMs, then track orders by hand.
What I built
Multi-tenant commerce SaaS: every shop gets a storefront on its own subdomain or domain, and reel comments, posts and influencer links turn into tracked orders with one stock pool, COD confirmation, referrals, campaigns, plans and background workers.
Engineering decisions
- Every third-party integration (WhatsApp, payments, shipping, OTP, storage) sits behind an adapter that defaults to a mock, so the whole product runs locally and going live is a config change.
- Environment config is validated at startup and the app refuses to boot when it is invalid, so a mock can never reach production by accident.
- Keyword rules classify reel comments first. A small model only sees the comments the rules miss.
- Background work such as cancelling unconfirmed COD orders and abandoned-cart reminders runs from a Postgres-backed queue, as a long-lived worker or a cron-driven tick, so the same code fits a VPS or a serverless host.
Stack
- TypeScript
- PostgreSQL
- Multi-tenant
- Background jobs
Written up in detail
- How to Take an AI-Built App from Prototype to Production
A working checklist for apps built with AI coding tools: secrets, data access, config, payments, background work, deploys and monitoring.
- Build the SaaS before you have the API keys
How Pavly runs every third-party integration behind adapters that default to mocks, and why keyword rules run before the AI.
Need something like this? See AI Product Rescue & Development and AI Integration Engineering and AI Agent Development and AI Deployment & Infrastructure.
More case studies
Have an AI product to ship?
Tell me what needs to be built, fixed, integrated or deployed. You work with me directly, the engineer who reads the code and ships it.