The catalan is very theoretical because it's extremly popular, but it's actually an opening you can play at the lower level with less amounts of theory knowlegde. Compared to extreamly sharp lines of the KID, usually you can get a decent position as white in the Catalan with a small edge and push for a long time even if your oponent takes you completely out of book early. Especially at a lower level a position top GMs see as completely drawn and equal can still be very playable since the opponent (and yourself) will definetly make mistakes. Many of the ideas in typical catalan positions are strong even with slight variables, such as a well timed Ne5 with pressure against the b pawn (if he doesn't fiancetto on the queenside), sacking the c pawn for pressure on the long diagonal, trying to push e4, putting the queenside pawns on a3 and b4 to stop c5 ect. The critical idea for black seems to be the Ra7 idea putting the queen on A8 to contest the diagonal but only yesterday Caruana outplayed Karjakin with a strong novelty in Tata Steel in this line, the game is fully annotated on chess base news if you wish to study it. With most other lines play is more quiet and even without theory you can find decent moves most of the time.
At least that's my experience from using the Catalan as a means of avoiding the extreamly theoretical Nimzo indian both online and otb. Sorry that I didn't answer your question, but as far as I know the least theoretical mainline opening as white in the d4 complex is the Catalan. Unless you want to play a system opening, in which case you could go for the london, the tromp, colle or something along those lines. Definetly playable but expect to see equality out of the opening.
Against these defences the most popular way to play seems to be to transpose into the king's indian fianchetto variation and the Catalan. But these openings can become very theoretical. Are there some simpler alternatives?