How much water you need
The Institute of Medicine’s reference value is 3.7 liters a day for men and 2.7 liters for women. That figure is total water including the water in your food, so the share expected to come from drinks is about 3.0 liters (13 cups) and 2.2 liters (9 cups). Two things about that number are worth more than the number itself.
The first is that it includes food. Quoting 3.7 liters as an amount to drink overstates it by roughly 0.7 liters, which is most of the gap between the figure you have probably seen and the one you actually need from a glass.
The second is where it came from. The panel set it at the median of what Americans already drank, and the reason they could do that is buried in the report: blood concentration was nearly identical in people in the lowest tenth of intake and people in the highest. The people drinking the least were not dehydrated. So the value describes normal intake rather than establishing a floor, and it is an Adequate Intake, the category the National Academies use precisely when the evidence cannot support a requirement.
What this tool used to do, and why it changed
Until August 2026 this page took your body weight, multiplied it by 35 mL, added fluid for exercise, added another 500 mL if you were on a GLP-1, and presented the result as your daily target. None of those three numbers came from the reference report or from anything else in our evidence base. We could not find a source for them because there is not one.
That is worth saying plainly, because the version you will find on most other sites still works that way. A personalized-looking number is more satisfying than a flat reference value. It is also, here, made up.
Everything counts, not just water
All beverages count, including coffee and tea. Their mild diuretic effect is far too small to cancel out the fluid they provide. Water-rich foods count too, and in the headline figure above they already are counted: fruit, vegetables, soup and yogurt supply a meaningful share of the total. If you eat a lot of produce, less of your total has to come from drinks.
Thirst, and urine that is pale rather than dark, remain the practical signals. They were the signals before anyone published a number and they still work.
If you are on a GLP-1
There is a reasonable argument that GLP-1 users need more from drinks: you are eating less, so less of your water arrives in food, so the same total has to come from somewhere. That argument follows from how the reference value is constructed. It has not been tested in anyone taking these medications, the report predates them by two decades, and we are not going to attach a number to an untested inference.
What we can say without inventing anything: persistent vomiting or diarrhea on a GLP-1 is a reason to contact your prescriber, not a reason to recalculate a hydration target.
A note on overdoing it
More is not infinitely better. Drinking very large volumes in a short window can dilute your blood sodium, a condition called hyponatremia that is uncommon but can be serious. If you have heart failure or kidney disease, or you have been told to limit fluid, your target comes from your doctor and not from a reference table. This page is educational, not medical advice.