Sort your reference into warm and cool masses and move the split until it matches what you see. Load a photo below — it never leaves your device.
Two passages can sit at exactly the same value and still fight each other, because one is warm where the light says it should be cool. Squinting finds value errors. It does not find temperature errors, because squinting removes the very information you would need.
A temperature map removes everything else instead. Each area becomes one of two colours, and the arrangement of warm against cool stands on its own — the same trick a value study plays with light and dark.
A neutral has no temperature. Forcing it into a camp would make every shadow in a low-saturation photo read as warm, and the map would show a pattern that is not in the picture.
So anything below a small saturation threshold keeps its own lightness in grey. If a large part of your reference stays grey, that is information: the subject has little colour to work with, and temperature is not the lever that will fix it. Reach for value contrast instead.
In daylight, warm light usually leaves cool shadows, because the shadow is lit by the sky rather than the sun. Reverse it and the scene reads as artificial light, or as a mistake.
It is a pattern, not a law. A warm wall bouncing light into a shadow breaks it honestly, and so does an overcast sky. Use the map to find out which case you are painting, then decide on purpose.
It sorts every part of your reference into warm or cool and paints each group a single colour, so the pattern of warm and cool masses becomes visible on its own. Areas with no real colour stay grey, because a neutral has no temperature to report.
Value tells you where the light falls; temperature tells you what kind of light it is. Two passages can sit at the same value and still read wrongly because a shadow is warm where the light source says it should be cool. Value and temperature are separate checks, and a painting can pass one and fail the other.
In daylight, warm light usually produces cool shadows, and cool light produces warm shadows. It follows from the shadow being lit by something other than the main source, most often the sky. Treat it as a pattern to check against rather than a law: a warm wall bouncing light into a shadow breaks it, and so does an overcast day.
Start in the middle and move it until the split matches what you see when you compare two passages directly. There is no correct number, because it depends on the light in your reference. The slider exists so you can find the line rather than accept one.
No. The map is computed in your browser on your own device. The image never leaves your computer.