The future of CubeCoders, and AMP Feature Freeze
A community letter from Michael, Director of CubeCoders.
“…just putting in what everyone wants now is easy, building something that is sustainable and reliable in the long term is hard and that is why this is necessary. It will pay off in the long run.”
Tl;dr - The short version:
- There’s an absolutely huge number of AMP users now, CubeCoders has been growing rapidly.
- AMP is entering a feature freeze in June, expected to last about 3 months.
- This does not affect the development of new game templates that run within AMP’s existing feature set.
- The feature freeze time will be used to work on the highest priority bugs and to establish new development processes.
- The focus will shift towards stability and maintainability in the short term, even if it means temporarily not being able to support certain new games that would require changes to AMP.
- We’re changing how releases and updates work so you have more control over how and when you update AMP.
- Older versions of AMP/ampinstmgr will be made available so you can keep on what works for you and upgrade at your own pace.
- We’re changing how support works to streamline the process.
- Staffing and pricing changes are on the cards (we remain committed to the pay-once model)
Moving Forward
CubeCoders is very much beginning to reach the point in its life where its own growth has started to become painful. What started off as a panel used by a few hundred people that supported about 20 different games has grown into something that supports well over 200 different applications. The number of users has also grown exponentially. Between the start of 2025 and today, more people have purchased and used AMP than in the whole of the preceding 10 years - complete with more demand for different games and applications, features, and requests for technical support.
One of the real challenges businesses face as they grow is quality. In our case we face a lot of demand for new features, for new games, and facilities that AMP needs to provide to accommodate those games. But one of the problems of constantly adding new features is that it’s very easy for quality to slip. Changing one component of the software to accommodate feature X can easily cause feature Y to misbehave or stop working in ways that aren’t easy to anticipate or test for. This is before trying to deal with the multitude of Linux distributions each with different tooling, packages, etc.
And so, we’re forced to ask the question. What is necessary to ensure the future of both CubeCoders and of AMP in the coming future, and how do we make sure that we maintain a quality product that is an asset to you and your communities.
In many ways, AMP is in a pretty good place right now - all of the main core functionality essentially works, but often with caveats. OIDC works, but perhaps not with your specific chosen provider. Backups work, but perhaps not as neatly as it should with the specific game you’re running. AMP covers such a broad feature set that it has far more of these ‘quirks’ than software we’re charging for really should, and regressions are more frequent than they should be.
When a company as large as Microsoft can’t ship Windows updates without breaking things, it’s almost certainly arrogant of us to assume that the same won’t happen to AMP periodically so we need to give users more tools and put in better processes to make sure that these events are minimised.
What is happening to AMP?
Starting at the beginning of June, AMP is officially going to enter a temporary feature freeze. No new functionality will be added to AMP, even if this means that support for a particular game has to be held back for a while if it is dependent on new features. We anticipate this will continue until September of this year.
During this time we’ll be doing more than just hunting down bugs, we’ll be putting in an entirely new set of actual processes and changes to help improve quality and more importantly make sure that users have a reliably, robust platform they can depend on.
As part of that, we’re going to be making the following changes:
- The ‘Match ADS version’ option will no longer be licence gated. Everyone will have the option to remain on their current AMP version and not be forced to upgrade when they create a new instance.
- We’re going to start allowing version picking, so if you’re setting up new AMP installations (say you have multiple nodes) - it’s easy to make sure they’re all on the same version even if you don’t want to use the latest version just yet. So older versions will be generally more available.
- The LTS branch is going to be brought back - so there will be a long term, stable version of AMP to which only bug fixes will be applied. This will have a much slower release cadence for actual feature updates but will still be able to get individual fixes as needed.
- The repositories will keep hold of historic ampinstmgr packages.
- Other changes are going to be put in place to make older AMP versions easily accessible when needed including the system-level tooling.
We’re also changing our internal processes to limit the impact of issues when they do happen. Commits will be restricted to a single concern at a time so that it’s possible to cherry-pick specific changes for LTS builds for example, and also makes specific changes easy to roll-back if they end up having unintended side effects without this also affecting other changes that may need to be kept.
What is happening to CubeCoders?
CubeCoders is starting to feel the strain of the size of its user base relative to the size of the company. So some of the changes moving forward are going to be designed to eliminate that strain.
Parts of the support process are going to become much more formalised and rigid in structure to make sure our attention is focused where it needs to be. Forum posts are going to give way to a much more guided process to make sure that engineer time isn’t being used inefficiently by them asking trivial follow-up questions or for logs that would be needed to diagnose the issue.
As mentioned above, some of the development processes are also being formalised into something more disciplined to make sure time is wasted less on managing clusters of unrelated changes.
James (IceOfWraith) has at this point been working with CubeCoders full-time since February, and AMP has massively benefited from the extra development time, and CubeCoders has benefited from having me being able to focus more on business aspects. Part of the plan for the second half of the year and early 2027 is to continue to bring on more engineers to manage the development and the support load.
The question that comes with hiring more people is of course, how to pay for it. We started charging £7.50/$10 USD for the cheapest version of AMP in 2019. If that price had increased with inflation it would be approximately £10/$13 but we’ve held the low price for 6 years despite the rising cost of goods and equipment and the cost of additional staff.
We remain very much committed to the pay-once model of software when it’s not being used commercially, and we want to ensure that AMP remains financially available to as large an audience as possible and that we charge a fair and reasonable price that is proportional to the value it offers. Right now, even the biggest consumer version of AMP (Advanced Edition) is still only the same price as a copy of Minecraft: Java Edition, and cheaper than a single copy of Rust - a 7 year old game. In fact it’s cheaper than almost all of the top titles that AMP is capable of setting up servers for.
So reviewing the pricing structure is very much on the cards, but rest assured that we will always make sure that AMPs pricing is fair, reasonable and accessible. If we do make a change to pricing, we will make sure that it’s announced ahead of time so that we can make sure we engage with the community over how this will affect them.
Closing
I really would like to thank every single one of you who uses AMP. CubeCoders is at this point a 13 year old company, what started off as me making a little Minecraft admin panel for my own use has grown into a massive product used by hundreds of thousands of people. I’d like to thank those of you who have been using AMP a long time and have stuck with it as it has grown, and I’d like to thank those of you who are new to AMP for your show of faith in adopting a new product for your servers.
It’s going to feel like a fairly quiet few months for many of you, and I appreciate that there will be some frustration if a feature freeze means your favourite feature doesn’t make it into AMP as quickly as you’d like, but just putting in what everyone wants now is easy, building something that is sustainable and reliable in the long term is hard and that is why this is necessary. It will pay off in the long run.