This amounts to nothing beyond anecdotal evidence. Unless the truth is absolute, it is useless. For instance, here is a list of reasons explaining why it's not done:
@4ward animations claims he is able to use GML, but refused to provide even the most basic of examples as to how he is.
Anecdotal evidence is not reliable from someone else who didn't claim to know what GML even is. While you didn't say or you did is of no relevance either, since that is something to open with when describing how you can tell others.
No proof
Pics or it didn't happen: No pics, no proof
No videos: same with pics. Not necessary, but not bad to have either
Word of mouth is only reliable with trusted/long-lasting members whom are reputable for telling what they know as the truth.
No full topic of it's own whatsoever. Less, if no credibility at best.
No plans to even release, 4ward even claimed that it even wouldn't let someone see if he had completed it either.
This reduces value in the whole plan as a whole, which makes what you plan to do hardly relevant anyways. Anybody would easily plan to make tools available to all, and to be anti-progressive is simply disregarding why they are making it in the first place.
More on the first point, why make an online mine-imator build if you didn't plan to release it to the public?
Why it's not possible/not going to work well:
Engine limitations
Gamemaker can perform online functions, but it cannot tank a whole 3d rendered engine like Mine-imator.
Online functionality is shoddy at best, even for gamemaker. A good example is trying to play Risk of Rain without issues.
Development time would take ages, especially for one person.
Easily replaced by tossing a file back and fourth, such as using Git or Dropbox.
@Emunator, who is an acclaimed moderator and well known GML modder for Mine-imator, claims the engine has trouble with latency while using online.
This verifies that Emunator has created a successful variant of Mine-imator Online, but it didn't work as well as he had hoped.
How it can work, if executed:
No real-time editing. Use periodic file updates.
This would be much easier on users, and reduces the need of having fast ping.
Reduces strain on the engine.
Prevent animation griefing
Option to user-lock things made by you. (Such as you animating a character, and nobody can make changes but you.)
Dedicated servers, with dropbox/git linking to projects.
I think the benefit outweighs the need to host LAN.
Having a single git/dropbox folder that updates with each file means anyone can simply access whatever's in the project at any time.
Limit 4 users per server, if able.
But honestly, using Git or Dropbox with friends and simply taking turns on a single project works much smoother than simply needing online function. I have a clear foundation on how it could work, but not considering any of what I've said means you are amounting to nothing with your efforts. I don't aim to be condescending when I say it, but you need to not try this if even those way above your pay grade have tried and failed. I'm sorry to say that it's merely a fantasy right now.