Graduate
Your spreadsheet, graduated into a real app.
A concept that turns the Excel workbook a small business runs on into a working app: it detects the tables, columns, data types and relationships, lets you review that structure, then generates forms, lists, filters, roles, an audit log and CSV/Excel export.
- Services
- Custom software
- System design
- Product design
- Technology
- Next.js
- TypeScript
- Python
- PostgreSQL
- Prisma
Graduate reading an orders spreadsheet: the detected table is outlined, title and totals rows are skipped, and each column gets a type
1 / 6The challenge
Many small businesses run on one shared Excel file: orders on one sheet, customers on another, with title rows, totals, typos and a “please don’t sort” note. It works until two people edit it at once, a formula breaks, or nobody can say who changed a price.
The approach
Graduate reads the workbook and builds a data model from it. It finds each table inside a sheet, skips titles and totals, infers column types, and matches values across sheets to suggest relationships. The owner reviews and corrects that structure, with every guess and its confidence shown, and only then is the app generated, so its forms, rules and permissions come from a model the business has confirmed.
Designed around
- Table, column and type detection from .xlsx files
- Relationship suggestions with match rates
- A review step before anything is generated
- Generated forms, lists, filters and saved views
- Roles with record- and field-level access
- An audit log that traces back to the original spreadsheet row
- CSV and Excel export of any view
