Case study

Audra

Audit, compliance and staff-testing platform for multi-store restaurant groups.

Status: Shipped
Year: 2026
Role: Solo builder (product, design, build and security), with coding agents
Users: 11 stores · 4 brands · 150+ staff tested monthly
Skills: Product scoping, Row-level security, POPIA, Auth, Testing UX

TL;DR

Problem: Changing an audit took about a week of waiting on developers. The app was Android-only while most staff used iPhones, losing signal meant losing your progress, and too few staff finished their monthly tests.
What I built: An alternative built from the floor up, running in the browser on any phone: audits the business edits itself, progress kept when signal drops, monthly staff tests, reporting that is easy to see, and equipment service tracking.
Result: Audit changes in hours, not a week. At least 90% staff test completion from month one, a considerable increase. Trialled at 2 stores, rolled out to 8, now 11 stores and 4 brands across South Africa, with 150+ staff tested monthly.
Stack: Next.js, React, TypeScript, Supabase, PostgreSQL, Vercel, Resend, Sentry
My role: Solo builder (product, design, build and security), with coding agents

Sample data throughout. The screens are the real app.

Audits
Paper checklists go missing.
Run audits on any phone.
Every site checked to the same standard.
Audits
A tick in a box proves nothing.
A fail needs a note, and asks for a photo.
You see what was wrong, not just that it was.
Open flags
Problems get found, then forgotten.
Every fail stays open until it is fixed.
Issues get closed, and you see how long each one waited.
Staff tests
Training is hard to prove.
Staff tests from a link.
You know who knows the menu, store by store.
Equipment
Nobody remembers the last service.
Every machine keeps its own history.
Services happen on time, not after a breakdown.
Reports
Reporting eats a day of spreadsheets.
Reports across every site.
Spot the site that needs help, at a glance.