Mitsubishi has the lions share of total numbers of contractors & support in our area, followed by Fujitsu, with the rest "in the statisitical noise". Contractors that have a lot of experience with Mitsubishi are Chaves in Hudson, Advanced Energy Conceptsin Fitchburg, and Meacham in Charleton, all of whom are on the "quality" end of installers, but it's still best to detail where/how to mount the outdoor unit.
I'm not sure who the better Fujitsu installers are in the neighborhood, but you can find some here. Daikin makes some of the better equipment (and are one of the bigger operators worldwide), but they don't have very many certified installers in central MA. If you stick with those three you'd be in pretty good shape from an equipment performance & reliability point of view.
In our climate it's not absolutely necessary to go with the "cold weather" series, units, but there are fewer maintenance/performance issues to worry about if you do. In the Mitsubishi lineup those would be the "Hyper Heating" sometimes called "H2i" series (MUZ- FExxNA). In Fujitsu it would Halcyon XLT-H series (AOU-xxRLS2-H). A typical "pretty-good" mini-split like those would have about 15,000 BTU/hr output at +5F per "ton" of rated cooling. (So a 12,000 BTU/hr = 1-ton unit is really 15,000 BTU/hr from a space heating perspective, in our climate.)
The layout of your house makes a difference in how many heads you'd need, and what size. Make a spreadsheet (Excel, etc) for calculating an I=B=R method heat load calculation on a room-by-room basis. You'll have to come up with U-factors (heat loss per square foot per degree) for your wall construction, windows, doors, attic, foundation (above grade only), etc. and the appropriate indoor/outdoor design temps. (Use the 99% outdoor design temp for your area if listed, or interpolate for the nearby listed locations. In Worcester the 99% design temp is +5F, and yes, it drops to -5F every few years, but it doesn't dwell at those temps long enough to really matter from a heating system point of view.) The basic formula for calculating the heat load is:
U-factor x area x temperature difference. So for each room you would have separate U-factors and areas for wall, window, ceiling to be calculated separately, then summed.
With a room-by-room heat load and a whole-house load I can steer you toward potential solutions.
Open common areas and adjacent rooms that open up directly onto it are usually candidates for running as one "zone", with a single head. A big doored-off room with lots of windows down at the end of the hall not so much. If a room doesn't have a heat load of at least 6-7000BTU/hr it's not a candidate for having it's own head, but there are "mini-duct" cassettes where you can split output between two nearby rooms with very short duct runs. Like space heaters and wood-stoves it's point-source heating, but it's not bad. Since they modulate rather than cycle on/off you get very stable room temps, and most of the time they're quieter than your refrigerator.
Since you have a heating fuel history on the place you can get a pretty good handle on the whole-house load by using propane-use against heating degree days. To do that we would need a zip code (to find a degreedays.net weatherstation nearest you), the # of gallons of a late-winter fill-up and the exact dates between that fill-up and the prior one, and the nameplate efficiency on your Rinnai heaters.
The picture of the drainwater heat exchanger (a 4" x 48" PowerPipe) is clipped from a Natural Resources Canada website. It looks like a Rinnai RL75i or similar, probably not set up for space heating, only hot water (since it would violate the Rinnai warranty to use it as a hydronic boiler.)