One business system across three interfaces
Yerias Hardware Store represents a business scenario with connected operations: products and categories, suppliers, purchasing, inventory, sales, receipts and users. The challenge was to bring the same business information to a web interface and a desktop point of sale.
Users and scope
The system serves sales staff, warehouse teams, administrators and users with different roles. The web client supports management and reporting; the Windows Forms application handles point-of-sale workflows and local operations.
Modules reflecting the business workflow
Catalog and inventory
Products, categories, brands, suppliers and stock control to keep operations organized.
Purchasing and sales
Purchase recording, cart/point-of-sale workflows and itemized receipts.
Users and roles
Authentication and authorization to separate responsibilities within the business solution.
Reports
Sales queries and reports, including date-range filters in the desktop application.
Web and desktop integration around the business domain
The ecosystem includes a .NET API for the React client and a WinForms application working with Entity Framework and SQL Server. This integration difference matters: the desktop application is not presented as just another HTTP frontend.
React Web Client React + TypeScript + MUI │ └── Axios → .NET 5 Web API │ ├── EF Core / JWT / Swagger │ └── SQL Server Windows Forms POS .NET Framework 4.7.2 + EF6 │ └── local access to the relational model and sales operations
Integration decisions and lessons
- Separating the API and web client supports a client-server architecture with Swagger documentation.
- The entity model and migrations help represent related business operations and make them reportable.
- The desktop application serves the local point-of-sale scenario, where the operational experience differs from the web dashboard.
- This case shows the cost of keeping logic coupled to different interfaces and provides a foundation for separating the domain and services later.
My contribution
I built and connected the sales ecosystem, working on the backend, web client and desktop application as parts of one business solution.
AI assistance alongside design work
AI supported requirements analysis, component exploration, debugging and documentation. Business logic, entity modeling, client integration and workflow validation required my own technical review and testing.
Translating business needs
The work starts with concrete operations—purchasing, selling, tracking stock and reporting—before choosing an interface.
State of the evidence
I do not attribute a specific public prompt or code-generation percentage to this repository; I describe it as an assisted development workflow.
Evidence for exploring the system
- The API documents endpoints, entities, JWT security and Swagger access.
- The web client contains catalog, sales, purchasing, user and reporting modules.
- The desktop repository contains login, product, sales, search and report forms.
- There is currently no deployed public demo. Recommended visual evidence includes the dashboard, point-of-sale workflow, a report and the architecture diagram.
- The system must run locally with SQL Server and the corresponding .NET tools.
Making deployment easier
The next improvement would be to extract domain logic into decoupled services, add automated tests, upgrade the API to a modern .NET version and prepare a controlled demo with sample data, so the product can be shown without a local SQL Server installation.