Could it be they’re both right? I mean, it’s possible that if you start rotating the model a few degrees on the z-axis, the measured dimensions on the platform might change.
Or, if it’s not a resin printer, you could always try the quick & dirty way: slice & start the print, pause it right after it extrudes the outline of the first layer (before it starts filling it in), and then measure the tolerances; resume if accurate, cancel if it’s off.
Most phone apps likely won’t reach the quality of a standalone scanner for a variety of reasons. For instance, a lot of the ones I’ve played with don’t optimize the shutter speed, which can wreck havoc when some captures contain motion blur.
That said, have you tried the free 3d Scanner App yet for iOS? Yes, you may need to clean up your scans like any other phone app, but since it also lets you use the front-facing TrueDepth sensor as well as the rear, it can return a pretty darn good scan when it can combine that TrueDepth data with the usual photogrammetry point cloud. Plus, it’s got a bunch of post-processing tools to modify your scans afterwards, like cropping out extraneous elements using a box or a sphere, selecting mesh pieces to remove/retain, etc. It’ll also export to SketchFab as well as your usual 3d printable file formats.
And no, I’m not affiliated with it; I’m just a happy user who was able to successfully use it to 3d print my doggo sleeping as well as scan some gallery exhibits.
Never built my own but hear me out, I have years of experience with a Flashforge Dreamer, which falls under #1. I personally found the second nozzle a hinderance to even 1 color prints since it would often bump into partial prints if/when the printed layer height is raised even slightly higher than it should be. The common workaround to reduce that from happening is by raising the nozzle about 0.5-1mm away from the bed, but of course that then causes additional issues.
Don’t get me wrong, the Dreamer’s a helluva machine, but considering how rarely I had something I want to print (or could print) in 2 colors, I found the additional nozzle a waste. Eventually I raised up the hot end entirely, conceding it wasn’t worth the trouble, and I also removed the second stepper motor entirely to reduce weight & increase fidelity.
Don’t let me keep you from tinkering if your mind’s set on building one, but if reliability is the name of the game, I’d look for an existing solution like a kit or predefined plans and build your own design off of that.