Taxi53

OMAps, Offline-first taxi, Innovation

Visit Website
May 7, 2026 I built a ride-hailing app that works without full internet. 20,000 rides later...

The problem

In Cuba, internet is expensive, slow, and unreliable. A taxi driver can spend 30-50% of their daily income just on mobile data.

Apps like Uber don't work offline. The local state-run app ("55") uses Google Maps and crashes the moment you lose signal.

So passengers in rural areas or during blackouts? They can't get a taxi at all.

The solution

I built the opposite.

Taxi53 stores maps directly on the device (Organic Maps, an open-source project). Routes are calculated locally (OSRM, another open-source engine). The only thing that travels over the network is a tiny sync signal: less than 1 MB per hour.

The server costs $20/month. No Google Maps API bill. No CDN for map tiles.

The test

We launched during Cuba's fuel crisis. No fuel for cars. No stable internet. No Google Play (Cuban developers can't publish).

The results:

  • 20,000+ real rides completed

  • 6,000+ downloads on a local APK site

  • Revenue: ~$4,000,000 CUP (Jul-Dec 2025) in a collapsed economy

  • Award: Winner, INNOVA 2025 (Cuban Ministry of Communications)

It worked when nothing else did.

Why I'm sharing this

I'm not building for Silicon Valley. I'm building for the 5 billion people who live with slow, expensive, or no internet.

I don't have a co-founder. I don't have VC funding. I just have a working product and 20,000 rides of proof.

If you're building for the rest of the world, I'd love to connect.

Technical details (for the curious)

  • Maps: Organic Maps (offline, stripped to 60-200MB per region)

  • Routing: OSRM (local, 0ms latency)

  • Backend: PostgreSQL + PostGIS, $20/month VPS

  • Sync: HTTPS every 10 seconds, <1 MB/hour

  • Stack: Android (Kotlin), Python/Go for backend

Questions for this community

  1. Has anyone here launched a marketplace (ride-hailing, delivery, etc.) in a low-connectivity market?

  2. Any advice on finding a business co-founder? I'm a technical founder.

  3. What's the best way to test Indonesia without moving there first?

Link to the dev.to article (more technical):
[ https://dev.to/leonardo_tq_13f83601e8513/20000-taxi-rides-in-cuba-what-i-learned-building-on-organic-maps-3k9b ]

Comment

About

In Cuba, internet is expensive. Drivers spend 50% of income on data. Uber fails offline. So I built Taxi53: maps on device, routes local, sync <1MB/hour. 20,000 rides tested. It works. That's why it exists.