Case Study · 2026
Visitor Tracker (Barack) — Visitor Management
Visitor Management
Multi-tenant mobile and SaaS solution for corporate front-desk visitor management. Handles walk-in check-ins, automated ID document scanning and OCR extraction via camera, appointment scheduling with host and time-slot selection, and document pickup tracking, with full light and dark mode support.
- Platform
- Mobile
- Role
- Flutter Mobile Developer (Team Project)
- Duration
- 4 months
- Year
- 2026
- Type
- Client project
- Status
- In production
Tech Stack
- Flutter
- Dart
- GetX
- Document Scanner / OCR
- Django REST Framework
- PostgreSQL
- Celery
- Docker
Context
At many corporate front desks, visitors still sign a paper logbook: handwriting is slow and often illegible, nobody knows who is currently on site, and appointments are tracked separately by phone or email.
The Challenge
Garder l'humain maître de ce que la machine a lu : l'OCR pré-remplit la fiche, mais l'agent doit pouvoir repérer et corriger une erreur de lecture en un coup d'œil, sans ralentir la file d'attente.
Approach
- 1 step
I developed the Flutter mobile app for receptionists. Staff photograph a visitor's ID card, the backend extracts details via OCR and pre-fills the check-in form for quick human verification, then records the host, purpose, and check-out time.
- 2 step
The app also manages scheduled appointments against staff availability and tracks document pickups.
- 3 step
Backed by a multi-tenant Django REST API (PostgreSQL, Celery, Docker) with automatic JWT token refresh and full light/dark themes.
Key Features

Enregistrement par scan
La pièce d'identité est lue par OCR et la fiche se remplit seule.

Correction de l'identité
L'agent vérifie et corrige ce qui a été mal lu.

Pointage de sortie
L'accueil sait qui est encore dans les locaux.
Rendez-vous
Pris sur les créneaux libres du personnel, puis confirmés ou annulés.
Results
Visitor registration shifted from manual typing to a quick photo-and-confirm flow, giving security real-time visibility into who is on the premises.
It taught me how to design OCR capture experiences where humans stay effortlessly in control of machine-read data.
Concevoir un parcours de capture où l'humain garde la main sur ce que la machine a lu, et tenir une session sûre sur la durée (jetons rafraîchis automatiquement).








