Back to work
Client workIn production
EST Marketplace
A storefront and back office for equipment, materials and vehicles, sold by bank transfer with nationwide delivery.
- TEFY's role
- Design, engineering, data model, admin tooling
- Related service
- Build AI products
- Stack
- Next.js App Router, Supabase Postgres and Storage, server actions
- Timeframe
- Dates to be confirmed
The problem
ESTechnologies sells equipment, materials and vehicles across nine departments to buyers who pay by bank transfer. They needed a storefront buyers trust and a back office staff can run without a developer on call.
How it works
- 01Buyer browses departments, adds to cart, places an order and receives a reference
- 02Bank details are shown once, only to the browser that placed the order
- 03Buyer uploads a transfer receipt; staff match it against the bank
- 04Staff sign in to a "Today" view that lists only what needs a person: unpaid orders, receipts to check, callbacks, out-of-stock items
AI functionality
None. A deliberately conventional storefront; AI was not needed for the problem.
Engineering decisions
- Order totals are computed on the server from database prices, so a tampered cart cannot change what an order costs
- Row level security on every table; only the catalogue is readable with the public key
- Photos upload directly to storage via signed tickets so phone photos are never blocked by request-size limits
- Pages read from the database on every request because staff edit prices and stale data would cost more than a round trip
- The staff password is never stored; a scrypt digest and a signed session cookie are verified on every admin action
Outcomes
Outcome figures (orders processed, time saved in the back office) have not yet been measured and will be added once the client shares them.
Have a similar problem?
A fit call is the quickest way to find out whether the same approach applies. Related service: Build AI products. Next case study: StarSprout.


