Depth of Opening Study

I am trying to establish some sort of opening repertoire using Chessable. Even if I set it to "Priority Only" there are 157 variations to memorize. I spend almost all my chess time trying to get it all down. I need (want) to spend some time on endgames and middle games.

One remedy I thought of is to limit the depth of of each variation to 6 or 8 moves. After all, at my level (1500s), I rarely get very far into a game before I am out of book.

Any advice from anyone?

Move by move is not enough. Cutting the variations short is a bad idea. The final goal of opening study is to learn :

- how will I get *all* my pieces out and complete development. You can't stop your lines until you have a full, clear idea of that (but it doesn't need to be move by move).

- what can the opponent do to annoy me meanwhile. There may be tricky problems that will be hard to solve at the board and opening study is there to solve problems, not to create more difficulties for oneself.

- what are the early middlegame goals. It's no good to get all your pieces out and have zero idea how to use them next. You will get stuck in a complex early middlegame you don't understand and again, your opening study will create more problems than it solves.

I understand it's quite hard to do at the beginning but without that you will just waste your time on openings, you would be better off looking at middlegames and endgames. And the approach is still valid at 1800 level, even 2000 level and a bit above (at ~2200, transpositions, typical structures in depth study and precise move-orders become necessary).

You're limiting the variations the wrong way. You should know the mainlines really well - that's all anyone plays at lower levels anyway. While it may feel like bluff, you are actually learning what the optimal playing of the opening looks like instead of getting destroyed by side variations (good as they may be). It's a common mistake people fall into.

There are two types of ways to learn openings - you could call them "variation tree" vs "model games" and for me the "model games" idea wins hands down every time. With the model games idea you're doing not only the opening but getting the middle game and endgame down as well.

I'd recommend Joe Gallagher's King's Indian books and play the model games out on a real board - they're a lot of fun especially as there are many quick games (so no worrying about the middle and endgame), and Gallagher himself is the protagonist of many of them. Also, Gallagher writes the more obvious things that other writers often leave out.

So how do I determine which lines are the "main lines" and where does one find annotated model games for those lines? Even if I can find an annotated game, how do I understand the annotation at my level?

Learning an opening repertoire takes years. You can't do it all at once.
Most important thing to do first is be sure you're satisfied with your choice of lines. That takes quite a while. It requires experimenting.

Your idea of limiting the depth is good. Start out learning 6 moves deep. Take some time off, focus on the other parts of the game a bit... then come back and go deeper.

Focus more on the common lines first (refer to lichess database for statistics). Try to identify the bad moves and ignore them - unless they're really common. Trust yourself to respond to them on principle.

But lastly... a repertoire isn't necessarily something you must play with perfect precision in order to get value from it. In the process of trying to memorize it you also familiarize yourself with many of the patterns and ideas. The grab-bag of patterns and ideas will be available when you go to play the positions.

There's a certain point in the opening where the algorithm ends, your ability to calculate correct continuations ends, and abstract logic consisting of chess principles must take over. You need to have a mental narrative of the game. It's not merely enough to copy a repertoire. Watch some high level games commentary - most of the games still end up in brand new positions by move 10. Once you fully appreciate that it will change how you think about the opening.

Really, as you memorize, it's good to think about the ideas behind every move. And that requires you doing the thinking, considering alteratives and why something works vs. why it doesn't. 
Sam Shankland annotates every move in his repertoire trying to distill the ideas. It takes a long time to do this but you will learn the opening much better this way.
You might even differ in what you'd like to play vs. what the chessable course recommends. Consider your options.. base decisions on your own values.
I've bought a few chessable courses myself, but mainly extracted the ideas and I ended up differing from what was recommended in probably 80% of the lines.
For all you know the chessable author chose a line because he thought you may be too dumb to play another one.

For example... is your opponent shifting alot of pieces toward your kingside? Don't play moves that shift your pieces away from the kingside. That's a logical thing to do... it's something that makes sense to you. 
You should develop the repertoire alongside your own logical thought processes, as much as you can. 
This becomes especially relevant toward the end of the opening, as complexity is increasing.

If it's a really theoretical line - like one people just play 17 moves deep every game... well, you could memorize just the particular concrete continuation in alot of depth, but I'd really advise you to deviate earlier. Because you're walking through a minefield of your opponents novelties by going into a theoretical line really deep like that.

I think a better use of your time would be to play over master games in your opening. Instead of trying to memorize 157 variations, see how Kasparov, Fischer, Carlsen, and others played. You will begin to see the patterns and themes. These will come up again and again in their games. And you will start to recognize them as they occur in your games.