A delivery shaped around the whole system.
Build more than listings and checkout: a buyer’s payment has to move through a stateful property transaction, keep seller funds isolated, route commission correctly and remain explainable if a payment provider or follow-up action fails.
Cuesoft provides ongoing product design, backend engineering, infrastructure and database work on Google Cloud, including the escrow and transaction model and the reconciliation controls behind property payments.
Experience design is part of the delivery.
Escrow is visible on the transaction screen
The transaction screen separates the amount funded for escrow protection, the platform fee and the property price. A buyer sees the status, property, verified counterparty and a downloadable receipt beside the financial breakdown, so escrow is in view before anyone presses Pay.
A timeline guides the next safe action
The timeline follows the kind of transaction: an inspection, an agent sale or rental, or a developer purchase. It tells a user when a tour is confirmed, when documents are ready to review, when funds can be released, when a refund applies and where to raise a dispute.
Role-aware transaction views
Seekers, landlords, agents, developers and administrators each have dedicated transaction lists and detail views. The interface gives each party a clear picture of the same transaction without asking them to interpret an internal operations screen.
The technical architecture.
A stateful transaction with an explicit lifecycle
Each property transaction has a buyer, seller, property, payment reference, exact kobo amounts, acceptance deadline and explicit lifecycle. Only pending and in-progress transactions are allowed to represent money still held in escrow.
Exact amounts and separated responsibilities
Escrow balances are held per seller in integer kobo. Each transaction records the gross charge, the amount held for the seller and the commission separately, so the platform can account for each movement without deriving it from a screen total.
Serverless application delivery on GCP
The API and client are delivered through Firebase App Hosting on Google Cloud, with managed PostgreSQL and Valkey services providing the transactional and rate-limit layers.
Details that make the difference.
Commission is configured by transaction type
Inspections, sales and rentals each carry their own commission rate, set in platform settings, and every transaction records its commission separately from the amount held for the seller.
Funds are reconciled per seller
An admin reconciliation check compares each seller’s escrow balance with the sum of that seller’s still-held property transactions and flags any difference. It deliberately does not settle for a platform-wide total, because an over-credit for one seller can hide an under-credit for another.
The payment path is designed for interruption
Funding intents hold a payment reference, exact breakdown and expiry. The flow detects lost listing reservations, avoids double local funding, moves ambiguous provider outcomes into reconciliation or refund states, and retries post-commit effects safely.
One active sale or rental per property at a time
The database allows one active sale or rental per listing while the service uses idempotent intent recovery. This keeps concurrent buyers from creating two live money paths for the same property.
Spend less on idle infrastructure.
Scale-to-zero is a deliberate operating posture
The platform avoids paying for permanently idle application instances. Durable payment and reconciliation states make that possible: work can be resumed and made visible, so no step depends on a background action that runs continuously.
Managed data services reduce routine operations
Managed PostgreSQL and Valkey provide the transactional and shared rate-limit layers without a separate database-operations programme. The deployment keeps termination protection on the managed data services.
- Product and UI design
- Payment orchestration
- Escrow reconciliation
- Cloud infrastructure
- Database engineering
- Google Cloud
- Next.js
- React
- Node.js
- PostgreSQL
- Redis
- Firebase
- Google Cloud