White Label Fleet Tracking: A GPS Platform With No Server to Run

Case study

White Label Fleet Tracking: A GPS Platform With No Server to Run

All case studies

I built a white label GPS fleet tracking platform that lets a Tunisian reseller serve its own clients under its own brand, without running a single server. One Cloud Function handles the whole API, Firestore pushes live positions to the map, and database rules keep every client's fleet isolated. The reseller runs it in production today.

0
Servers to administer
1
Cloud Function for the whole API
80
Concurrent requests per instance
OsmAnd
Standard tracker protocol

The problem

A Tunisian reseller wanted to sell GPS fleet tracking to transport and delivery companies under its own brand. The classic way to build that product starts with infrastructure: a server running around the clock to ingest GPS frames, a websocket layer to push positions to the map, and a database sized for the worst case.

For a reseller starting with a handful of clients, that model breaks down quickly:

  • Fixed costs arrive before revenue. The servers cost the same whether ten vehicles are tracked or a thousand.
  • Someone has to operate it. Updates, monitoring and outages become a job the reseller never signed up for.
  • The brand is not theirs. Most ready made platforms show the vendor's name, not the reseller's.
  • Isolation is fragile. When several client fleets share one system, a single badly written query can expose another company's vehicles.

The approach

I took the server out of the equation and let managed cloud services carry the load, so that cost follows real usage instead of preceding it.

One function instead of a server

The whole REST API and the tracker ingestion endpoint run in a single Cloud Function on Express 5, configured with a concurrency of eighty. One instance absorbs a large number of vehicles before a second one ever starts, and nothing runs when nothing moves.

Live positions without a websocket layer

Instead of building and babysitting a push infrastructure, the map subscribes directly to Firestore listeners. A new position written to the database appears on the fleet manager's screen within moments. The interface is React with a Leaflet map over OpenStreetMap, installable as a progressive web app on a phone home screen.

White label and isolation by design

Brand and subdomain are tenant settings, so a reseller installs its identity without touching code. Firebase Auth custom claims carry the tenant and the role, and Firestore security rules enforce the separation at the database level. A bug in the interface cannot leak another client's fleet, because the database itself refuses the read.

No proprietary hardware

Trackers connect over the OsmAnd protocol, which smartphones and common GPS boxes already speak, so onboarding a vehicle is a configuration step rather than a hardware purchase. A scheduled function also animates a realistic demo fleet, so a prospect sees vehicles moving before a single tracker is installed.

Outcome

The reseller runs the platform in production today and serves its own fleet clients under its own brand.

  • No infrastructure to operate: there is no server to patch, monitor or restart.
  • Costs that scale with the business: the cloud bill grows with tracked vehicles, not ahead of them.
  • A faster sales cycle: the built in demo fleet shows a working product in the very first meeting.
  • Isolation that does not depend on perfect code: tenant separation is enforced by the database rules.

Onboarding a second reseller is now a tenant to create and a subdomain to point, not a deployment to rebuild. The full technical breakdown is on the white label fleet tracking project page.

Tech stack

React
TypeScript
Firebase
Cloud Functions
Firestore
Leaflet
OsmAnd
PWA

Related services