Tiwaz, I suppose you have not read the list of authors in Lords of men and the like, right? In general I would consider that someone like Timothy has a generally better grasp of the line and the intended setting and what the Code means than myself if I were to enter into an argument here
The setting has a fairly clear cut parameter that is mostly agreed by players. It is WAY less dark than the World of Darkness universe by a long shot. For starters, the code is fairly "modern" and more a set of directives than a letter of the rules (except the dealing with demons bit, that is literal) in most cases. Tribunals have a lot of liberality on how they interpret the clauses most of the time.
If you want to be gritty and dark it is perfectly fine (I have played sagas like that) but it is not the official setting.
This is entirely doable within even the vanilla Ars Magica setting - so much so that there are canon covenants that match each of these models - a covenant that is a bunch of villages in a valley sits in (I believe) the Transylvanian tribunal, and a covenant that is integrated into various buildings in a city is in the Rhine. Other, similar covenants are mentioned elsewhere too - and Fengheld in the Rhine probably takes a prize for most distributed covenant, technically spanning two tribunals.
There is no reason a covenant (in vanilla Ars Magica) couldn't consist of a hoplite or three and their train who are a mercenary company. As long as they don't go overtly using/selling magic in the service of a lord, it's a great way of moving around Mythic Europe.
The new grogs book gives some good suggestions on how to handle this - and for the kind of game you want to run I'd recommend picking it up. It's a good book.
A way to handle distributed magi like this is to tie companions to a specific magus, then have each magus-companion group supported by enough player grogs to make up numbers. Thus, in most cases your 'adventuring party' will be a magus, a companion and a few grogs. The lack of close communication between the members is less of an issue, because each sub-group is independent enough to run stories for the entire troupe (players) without having to drag every character into the same room all at once.
With a very distributed covenant like this, I would allow some form of easy-communication between sites - perhaps in the form of a magical item or similar that comes from the covenant build points. While such might fly in the face of your otherwise low-magic setting desires, if it binds the player characters together more then it's a win! And if it is near-unique in your setting, it just makes it cooler. It could even become something of a story hook in its own right.
Anyway, I still strongly recommend the covenant concept for the agreement-between-players to define the kinds of characters that will fit with the group. So much so that I'd even recommend such a concept for any RPG, not just Ars Magica. It gives much more versimilitude to a group to say 'you are all part of a mercenary company...' than 'you meet in a bar, and despite being total strangers of diametrically opposed philosophies you decide to go risk your lives together...'
And if the players have a hand in shaping the covenant, they have more invested in it - and in the saga as a whole.
FWIW, I agree with Timothy; this is my understanding of vanilla ArM5.
Magi can, do, and have, cast spectacular spell effects in public. Sometimes with Hermetic legal consequences, sometimes not, it all depends on the context of what was actually done.
I think for the "coven" model to work well, the grogs and companions certainly have to have reasons to interact. If the coven is based in a bunch of nearby villages some ideas for interaction include:
If only one village has a church, then the other villagers will then have to regularly travel to that village for religious festivals, funeral, marriages, and baptisms. Also, the village with a church may periodically be the site of a canon court, so other villagers will have to go there for court cases. Likewise, if the "coven" grogs are pagan have an important cult site at/near one of the villages.
Ensure that the villages interact economically. Maybe one is on a major river/road and so is the regional "port" for long distance commerce and will therefore probably have a market. Another village might have a monopoly on mills. Another might be based around a mine, or something. Basically make sure that at least a couple of the villages have unique things that the other villagers need access to.
Make sure there are family relationships that bridge the villages. Then you can have lots of family reasons to interact (a birthday celebration, feuds, arguments over wills, an ancient great-uncle that no-one wants to look after, childhood rivalries to extend into adulthood, etc) --- before you know it, you will have fodder for greek tragedies/comedies spanning the villages.
In general I would consider that someone like Timothy has a generally better grasp of the line and the intended setting and what the Code means than myself if I were to enter into an argument here