inTable
My role
I built the front end on my own, from November 2025 to May 2026, with practically all of the repository's 192 commits. The API runs on InBot's back end.
I built the front end of inTable, a table manager with corporate login (Keycloak) and permissions by company and department. Each table has typed columns, filters, CSV import and export (with asynchronous jobs for large tables) and schema export as JSON. Bots access the same tables through the API, with per-table keys and permission scopes. There are also dynamic dashboards built with Recharts, which can be edited by dragging and resizing the cards. Stack: React + TypeScript + Vite, TanStack Query and Zod.
Before
Every bot project that needed to store customer data solved it its own way, with no shared screen for customers to look up and fix their data.
The problem
InBot's bots need to read and write structured customer data, such as registrations, appointments, catalogs and survey answers. And customer teams need to view and manage that data without relying on a developer.
The solution
I built the front end of inTable, a table manager with corporate login (Keycloak) and permissions by company and department. Each table has typed columns, filters, CSV import and export (with asynchronous jobs for large tables) and schema export as JSON. Bots access the same tables through the API, with per-table keys and permission scopes. There are also dynamic dashboards built with Recharts, which can be edited by dragging and resizing the cards. Stack: React + TypeScript + Vite, TanStack Query and Zod.
Decisions and trade-offs
Permissions through Keycloak groups (company and department) instead of adding users table by table. Exports run as an asynchronous job so the screen doesn't freeze on large tables. Dashboards laid out on a grid, with each card placed by coordinates, so customers can rearrange them without code.
The results
In production since December 2025, inTable now holds about 7 million rows across 301 tables from 34 companies and 65 departments, and 190 tables were updated in the last 30 days. It became the foundation InBot's bots and AI skills use to read and write data. The front end has 178 unit tests.
Results by the numbers
- ~7 mirows stored
- 301tables in production
- 34companies using it
- 178unit tests