Why guessing at random settings wastes filament
First-layer problems have several possible causes that produce visually similar symptoms: bed adhesion, Z-offset, contamination, thermal drift, and even part geometry can all show up as a first layer that will not stick or looks wrong. Changing several settings at once might accidentally fix the problem, but you will not know which change actually worked, which means the next failure starts the guessing over again. This tool works the other way: from an observed symptom and a few context details, toward a small ranked list of the most likely causes and the specific next test for each one.
How the ranking works
Each symptom pattern (corners lifting, nothing sticking, a squished layer, or gaps between lines) points toward different underlying causes with different starting weights. Context you provide, a dirty bed surface, a confirmed-clean nozzle, or a drafty environment, shifts those weights up or down. A confirmed-clean nozzle, for example, removes contamination from serious contention even though it is a common general first-layer cause, since you have already ruled it out for this specific print. The result is the top two or three hypotheses left standing, each with one concrete next test.
Worked example
Corners lifting off the bed, bed recently cleaned, nozzle confirmed clean, normal room: the tool ranks bed adhesion (temperature or surface prep) and thermal/draft-related warping as the top two hypotheses, each with its own next test, and does not raise contamination as a leading cause since the nozzle is already known to be clean. This matches the general pattern that corner lifting on an otherwise clean setup usually traces back to adhesion or cooling-related warping rather than a dirty nozzle.
Limitations
This tool ranks plausible causes from what you describe; it cannot see your printer, so treat the list as a place to start testing, not a confirmed diagnosis. A first-layer problem can also come from a mechanical issue, a bent gantry, a failing bed-leveling sensor, a loose belt, that no software-side triage can detect from a text description alone.