Shells carry more of the load than most people expect
A common assumption is that raising infill density is the main way to strengthen a printed part. For many load cases that is backwards. The outer walls are continuous, well-adhered material running the full height of the part, which puts them directly in the load path for bending and tension. Infill is a repeating internal lattice that mainly resists being crushed locally; it contributes far less to resisting a part bending or being pulled apart than an equivalent volume of extra wall material does. This tool starts from that distinction and asks what your part's load actually does before recommending which lever to pull.
How the comparison works
The load type you select determines whether a wall-first or infill-first change is suggested as the best fit, based on how each construction resists that specific kind of load. Bending and tension favor extra wall count; compression and impact give infill density more credit, since impact benefits from both approaches together. The result also weighs your stated priority, since a wall increase is usually a smaller mass and time cost than an equivalent infill increase for a comparable strength gain, at least until infill density gets quite high.
Worked example
A functional bracket under bending load, currently 2 walls and 20% infill, weight priority: the tool recommends a wall-first change (raising to 4 walls, keeping infill near 20%) as the best fit, since bending is carried mainly by the outer shell and a wall increase adds less mass than a large infill jump. Raising infill toward 45% appears as the alternative, useful mainly if the load is actually more compressive than the bending description suggests.
Limitations
This is a qualitative comparison, not a strength or load calculation, and it never estimates a safe load or a factor of safety. Layer adhesion in the print direction is often the true weak point for a tension load, and no wall or infill change fixes a part loaded mainly across its layer lines; reorientation or a redesign may matter more than either lever here. Validate a load-bearing design with a real test part before relying on it.