Protomaps 4 min read

What Your Map API Bill Won’t Tell You About Protomaps

Key takeaways

  • Protomaps lets you store and serve map tiles from infrastructure you control.
  • Hosting tiles does not replace address search or route planning.
  • A fair cost comparison includes infrastructure, ongoing operations, and any remaining API fees.
  • Greater control over hosting brings responsibility for updates, deployment failures, and performance.

A growing product can turn a manageable map API bill into a reason to rethink your stack. Protomaps makes hosting your own map tiles an option. Whether that saves money depends on what you need the map to do—and who will keep it running.

First, decide which part of the map you’re replacing

Users see one map. Your infrastructure serves several distinct functions.

The background map is one. Turning an address into coordinates is another. Calculating a driving route is another still. They may share a screen, but they need separate entries in your cost calculation.

Protomaps uses PMTiles, a format that packages many map tiles into a single file. Tiles are small pieces of a larger map; the application loads the pieces needed for the current view. Because it can read the relevant portions of a PMTiles file, you can distribute that file through infrastructure such as object storage.

That covers map tile delivery. It does not automatically provide geocoding, which converts addresses into coordinates, or a routing service.

A store locator and a navigation app therefore have very different replacement costs. Before comparing prices, break down your existing bill by function. You need to know which charges self-hosting could actually remove.

The infrastructure savings can disappear into working hours

Self-hosting changes where the money goes. Instead of paying a provider to serve tiles, you pay for storage, file requests, and data transfer.

You also pay for someone’s time.

Consider a hypothetical monthly tile API bill of ₩1,000,000. If storage and transfer for a self-hosted setup cost ₩200,000, the apparent saving is ₩800,000—an 80% reduction.

Now add 10 hours of extra operations each month. At an assumed internal labor cost of ₩80,000 per hour, those hours cost another ₩800,000. Your total is back to ₩1,000,000.

That is an illustrative budget, not a Protomaps price quote. It shows why a smaller infrastructure invoice can leave total costs unchanged.

Those hours might go toward updating map data, checking deployments, or investigating slow loading. A team that already runs storage and delivery infrastructure may absorb the work with little additional effort. A team building those capabilities from scratch also faces setup and migration costs.

The useful comparison is the total cost of delivering the same functions at a similar level of service. Engineering time belongs in that calculation, even when it never appears on the cloud bill.

Portability has value, but someone owns the pager

Cost is only one reason to consider self-hosting.

When you control the map files and delivery setup, changing storage or delivery providers can become easier. That gives you more room to respond when a vendor changes its pricing. Portability can be valuable even when the immediate savings are modest.

Some dependencies remain. You still rely on a source of map data and a library that draws the map. If you retain a separate address search service, that relationship remains too.

The operational responsibilities also need names attached to them:

  • Who decides how frequently new roads appear in the map?
  • Who rolls back a failed data deployment?
  • Who investigates when maps become slow?

These are part of owning the service. So are data licensing and attribution requirements. Hosting a file yourself does not change the terms governing its underlying data.

Run a small pilot with real usage

Start with a feature inventory. Separate background maps and location markers from address autocomplete, geocoding, and routing. That defines the scope of what you can replace.

Then look at how people use the map. Repeatedly showing the same neighborhood creates a different request and transfer pattern from letting users pan and zoom around the world. Monthly visitor count alone cannot tell you what delivery will cost.

A limited rollout should measure:

  • File requests and data transfer during actual map use.
  • Costs and loading speed before and after caching.
  • Time spent updating data and deploying changes.
  • Fees for features that still require external APIs.

Use those measurements to compare the savings with the added work. Include the initial transition effort, and make sure someone has capacity to own ongoing operations.

Protomaps can make sense when lower delivery costs and greater control outweigh the work your team takes on. The deciding number is the cost of keeping the map useful and reliable—not just the amount printed on the next invoice.

Protomaps Map APIs Self-Hosting

Comments

    Loading comments...