I think you're trying to tackle too much at once, and you're over-complicating it. Sketch out what you are trying to do on paper first, making it as simple as you possibly can. Then, make it even simpler. Throw away any ideas that are 'nice to have' in the first version (the axiom You Ain't Gonna Need It is very helpful here).
Your first aim is to develop a prototype that just about works. Do not get bogged down in fancy UI or learning clever technology unless it is absolutely essential to the success of your enterprise.
Also, approach possible users, or people you know in the field of your project, and ask them whether they think it is a good idea. Be willing to hear that the scope of the project needs refining.
Then, set up Git on your local machine, and get something working. Don't worry about whether you are doing it perfectly at first. If you actually have existing team members, push it to a remote Git site so they can pull from there1. A VPS with SSH access is suitable for this, it'll cost you a few GBP/USD per month. Learn branching and merging if you can, it is very useful.
When you have stuff taking shape, push to a test machine on the web (re-using the same VPS) and try it out there. Let some users onto it as early as you can, and release as often as you can.
It is a good idea to write tests, yes, but don't spend excessive time worrying about CI and your bug reporting mechanims before your prototype is ready.
1 Always keep an offsite backup, of course.