Built by a Developer, for Every Developer
horizOn started as one developer's project and grew into a platform for every developer. Here's the honest story behind it.
Why horizOn Exists
I was working on a game project and needed a proper backend. Not just a database, but the kind of infrastructure the big providers offer: globally distributed servers with low latency in Europe, the US, and Asia. Database replication, automatic backups, the works.
The problem? Every provider that offers this level of infrastructure, like AWS, Firebase, or PlayFab, charges per use. API calls, data transfer, compute time, it all adds up. I never knew what my monthly bill would look like, especially if the game gained traction.
That didn't seem right. As an indie developer, I needed to plan my budget, not watch usage meters and hope for the best. So I decided to build it myself.
Over time, it turned into a full Backend as a Service: authentication, cloud saves, leaderboards, remote config, news, localization, gift codes, player feedback. I released SDKs for Unity, Godot and Unreal so other developers could use it too.
What began as one developer's tool is now a platform for every developer: a backend with servers in Europe, the US and Asia, available in 15 languages and run by ProjectMakers.
And horizOn is here to stay. ProjectMakers uses it in its own client projects, so the backend keeps running and keeps being maintained, no matter how many studios sign up. You are not building on an experiment, but on infrastructure we rely on ourselves every day.
Built and backed by ProjectMakers
horizOn is built and backed by ProjectMakers, a German sole proprietorship that Simon Daniel März has run since 2016 and that ships web, desktop and mobile apps and games. The same business that builds custom software for clients runs and funds horizOn, so the platform is not a side project. It rests on a real, working business with its own revenue.
ProjectMakers is owner-led and hands-on. Simon Daniel März still works directly on the product, which is why horizOn keeps its promise of clear, fixed pricing and reliable, well-built infrastructure.
Three people work on horizOn: Daniel as the owner and two freelance developers who work with ProjectMakers on a regular basis. Both developers also have access to the servers and the backups, so running horizOn does not depend on one person.
horizOn will not disappear overnight. Our own apps and games and our client projects run on horizOn, so we depend on it ourselves. Our Terms (clause 11.5) commit us to announce an end of horizOn at least 12 months ahead and to publish the backend as open source. In that time we will also extend simpleServer as your way to self-host.
Real Numbers, No Marketing Fluff
We believe in transparency. These are our actual numbers. Not projections, not targets.
Live values from the running system. Details at horizon.pm/status.
Infrastructure You Can Trust
Three regions, replicated data, daily backups.
Multi-Region
Servers in Europe, the United States, and Asia. Your players connect to the nearest endpoint for the lowest latency.
Replicated Databases
Your data is copied to three database nodes in three regions. If one node fails, the other two still hold it and keep serving reads.
Daily Backups
Every account database is backed up once a day. Daily backups are kept for 14 days, weekly backups for 8 weeks and monthly backups for 12 months. On request we restore a single account.
How Fixed Pricing Works
Most providers charge per API call, per GB transferred, per compute minute. We don't.
I had been burned by pay-per-use bills that punished success, so horizOn does the opposite: you pick a tier and the monthly price stays fixed, no matter how well your game does.
That is the whole idea behind fixed pricing. See exactly what each plan costs on the pricing page.
Pay-Per-Use Providers
horizOn
Quality Through Automated Testing
Every change is tested automatically before it reaches production. No shortcuts.
Frontend
Backend
We run our full test suite on every commit. Automated CI/CD pipelines ensure nothing reaches production without passing all checks.
Test cases counted in the repositories on 12 October 2026.
* Measured on our status page. Uptime: backend average over the last 90 days. The backend counts as down only when no region answers, because a load balancer sends requests to the regions that are still up. Response time: backend average over the last 14 days, measured inside each region, without the network path to the player. It is the mean of the daily averages across all regions, the status page shows the values per region. Availability target in our Terms and Conditions: 99.98% for the backend.