The app doesn't make sunrise direction predictions. It shows the location of the Sun over a Monopole model.
And you can then use those to determine where the sun should appear at a given location (i.e. the direction to the sun) and compare that to reality.
It fails repeatedly.
The direction of the sun in relation to the map is generally consistent with what is seen during the day, however.
No it doesn't.
The azimuth matches fairly well at mid day, but the elevation is wrong.
The azimuth matches quite poorly away from mid day and the elevation is still wrong.
It is generally inconsistent with what is actually observed.
The East/West sunrise/sunset direction is part of the same explanation for why the sun does not change size and is discussed at https://wiki.tfes.org/Equinox#A_Flat_Earth_Equinox
Which doesn't work at all. And doesn't have the equinox being special in any way.
It relies upon only a tiny portion of Earth being illuminated at any given time, rather than roughly half.
Where for example, you would only see the sun if it was above a location some 10s of km away.
But even then, it would still only appear due east for the people along the path of the sub-solar point.
For those north or south, it wouldn't. And it wouldn't appear to rise at all.
And of course, it would mean most regions would be in darkness for months, only getting sunlight for a brief period each day for a few days. Other regions would always be in darkness.
What you actually need to explain the due east sunrise (without appealing to magic) is for the sun to be in basically the same direction for all those along a line of longitude. That would mean that the sun is very far away and illuminating a very large region.
As I have shown elsewhere, simple trig shows this to be the case.
Lets consider a time where for one line of longitude it is roughly due east, and for the opposite line it is roughly due west. (Which does occur)
Lets look at the 2 points on the equator, and assume that it is 1 degree off.
So for a point 10000 km away from the centre it is 1 degree north of due east, and for the other point it is 1 degree north of due west.
This results in an isosceles triangle (the other point being the sun), with the equal angle being 89 degrees, and the base being 20 000 km.
This makes the height tan(89 deg)*10 000 km =~ 570 000 km.
That puts it well off Earth.
So no, that excuse of just illuminating a tiny region simply doesn't work.
Likewise, for the exact same problems, with the required region of illumination, it not changing size doesn't work.
Sure, if the sun was at an altitude of roughly 5000 km and was only visible when the sub-solar point was within 10 km, it would appear roughly the same size.
That would also mean it would only every appear high in the sky, basically straight up. Never rising or setting.
What you need is for the sun to appear to go near the horizon, i.e. rise and set. But if the sun remains at roughly 5000 km, this requires it to be very far away horizontally.
This then results in a massive difference in the distance to the sun and thus a massive difference in the apparent size.
You are quite happy using cars and jet skis to pretend that the sun should rise due east, but you then completely ignore that they start off tiny and change size dramatically.
So no actual answer there.
So again, we have FE completely unable to explain the apparent direction to the sun nor why it remains roughly the same size, nor why it appears to rise and set.
Meanwhile, we have a RE model which is capable of accurately predicting the apparent direction to the sun, and which explains why it remains roughly the same size and why it appears to rise and set.
So why should REers be accepting defeat? It sure seems like FEers should be.