America's favorite psychotropic — taken by almost everyone, measured by almost no one.
Everything on this page is an estimate built from published averages. It is a mirror, not a clinician. It starts empty: add what you've actually had today and the numbers above will follow you.
"A cup" is a terrible unit. Brewed coffee alone spans roughly 80 to 200 mg per 8 oz depending on bean, roast, grind and brew. Build your real day below.
I just had… (added at the current time; drag it on the curve to move it)
Or add anything, at any time
| Time | What | mg |
|---|
Values are typical per-serving amounts from FDA, USDA and manufacturer data; real cups vary. Guarana and "energy blends" are often not counted on labels.
FDA describes about 400 mg/day as an amount not generally associated with dangerous effects in healthy adults. That is a population statement, not a personal safe limit: pregnancy, anxiety disorders, arrhythmias, some medications and slow metabolism all move the line.
Drag any dose along the day and watch the whole curve move. This is the moment caffeine pharmacology stops being a rule ("nothing after noon") and becomes a mechanism.
The textbook average is about 5 hours, but the honest range across people is 2 to 12 hours, and a few things move it a lot. Tick what applies. The slider is an estimate, not CYP1A2 genotyping.
The 23andMe "Caffeine Consumption" report predicts how much coffee people with your variants tend to drink, from two markers that together explain about 1% of the difference between people. It doesn't tell you your clearance rate. The markers below are in your raw data anyway (Browse raw data → search the rs number); together they nudge the half-life estimate and the sleep projection. Each is a small effect. Pregnancy, fluvoxamine and the pill all matter more than any of these.
What the evidence supports, in one paragraph each. CYP1A2*1F: AA carriers induce the enzyme more readily, which is why the "fast metabolizer" label shows up in the cardiology literature (slow metabolizers had more heart attacks with heavy coffee in a 2006 case-control study) and in sports science (the ergogenic effect of caffeine was confined to AA in a 2018 trial). The catch: the variant mostly changes inducibility, so it separates fast from slow clearly in smokers and only modestly in non-smokers. rs2472297 and the AHR marker come from genome-wide studies of coffee drinking; a later study of blood caffeine metabolites showed the same loci change how fast caffeine is converted to paraxanthine, so faster clearers drink more to hold the same level. ADORA2A TT is the best-replicated "sensitive" genotype: more caffeine-induced anxiety at 150 mg, larger effects on sleep EEG, and lower habitual intake. It changes how much a milligram does, not how long it lasts. The 23andMe report's headline prediction is honest science with a small effect: a variant or two shifts predicted intake by a tenth of a cup a day.
The evidence people should know: in a randomized crossover trial, 400 mg taken 6 hours before bed cut objectively measured sleep by about an hour, while participants barely noticed on their own ratings. "I can fall asleep" is not the same as "caffeine isn't affecting my sleep." Meta-analytic estimates suggest a standard cup needs roughly 9 hours to clear enough to leave sleep unaffected, and a 400-mg dose about 13.
The loop to watch for: caffeine → lighter, shorter sleep → morning fatigue → more caffeine, earlier and larger.
One-compartment model with first-order absorption (absorption half-time about 15 min, so peak around 30–60 min) and first-order elimination at your estimated half-life. Doses stack additively. The simple memory version: caffeine remaining = dose × ½^(hours ÷ half-life), so at a 5-hour half-life 200 mg at 8 AM ≈ 100 mg at 1 PM ≈ 50 mg at 6 PM ≈ 25 mg at 11 PM. Amounts are total caffeine in the body, not plasma concentration, so weight doesn't change the number, only how it feels.
Most caffeine confusion collapses several distinct phenomena into one word. Pull them apart and the picture gets much clearer — and much less moralized.
| Concept | What it means | Caffeine example |
|---|---|---|
| Tolerance | Same dose, less effect. Develops within days of daily use; fades within a week or two off. | "One cup used to wake me up; now I need three." |
| Physiologic dependence | The nervous system has adapted (more adenosine receptors, roughly speaking); removing the drug produces withdrawal. | Skip coffee → headache, fatigue, irritability. |
| Psychological reliance | A learned belief that you need it to perform, cope, socialize, start work. | "I literally cannot write without coffee." |
| Habit | Cue → behavior, automatically, without a decision. | Wake up → walk to the espresso machine. |
| Craving | Subjective wanting or urge. | Thinking about coffee through a morning meeting. |
| Problematic use | Continued use despite consequences, or loss of control over amount. | "It gives me insomnia and panic but I can't cut down." |
| Use disorder | A broader syndrome: impaired control, continued use despite harm, withdrawal. | Much rarer than plain dependence. |
Caffeine withdrawal is a formal DSM-5-TR diagnosis. Caffeine use disorder is not; it sits in Section III as a "condition for further study," with proposed criteria that emphasize failed attempts to cut down, continued use despite harm, and use to avoid withdrawal.
You can be physiologically dependent on caffeine without having a caffeine addiction.
Not one "are you addicted?" score. Three independent domains, never summed, because someone can have obvious withdrawal and use caffeine perfectly rationally, while someone else has mild withdrawal and keeps using it through panic attacks and failed quit attempts.
Best guess is fine. If you can't remember your last full day without caffeine, answer from the last time it happened by accident: a travel day, a stomach bug, a fast, a morning the machine broke. If there hasn't been one, say so; that's an answer too.
These items were written for this page. The first dial paraphrases the DSM-5-TR caffeine withdrawal criteria; the third paraphrases the proposed caffeine use disorder criteria. They are not a validated instrument; if you want one, the Caffeine Withdrawal Symptom Questionnaire and the Caffeine Use Disorder Questionnaire exist in the research literature. Answer for the past three months.
Answer the questions to see your profile. Each dial scores 0–10 on its own.
Your card is drawn in your browser from your answers and your logged doses; nothing is uploaded.
Caffeine works mainly by blocking adenosine receptors, the brain's accumulating "sleep pressure" signal, not by pumping dopamine the way amphetamine does. Block those receptors every day and the brain compensates by upregulating them. Remove caffeine and adenosine now signals louder than baseline: headache (adenosine dilates cerebral vessels), fatigue, fog, low mood.
Blue dots are adenosine, the sleep-pressure signal. Amber blocks are caffeine sitting in the receptors. The bar is how loudly adenosine gets heard. Run the days forward, then switch caffeine off.
In daily users, a large share of the morning "lift" is withdrawal reversal: caffeine returning you to the baseline that yesterday's caffeine moved. Studies of habitual users show that the alertness gain after the first cup mostly restores performance to that of non-users, rather than exceeding it.
The experiment: rate your alertness (1–10) at 9 AM for five days as usual. Then taper to under 50 mg/day for 10 days and rate again. If the numbers converge, your morning coffee was mostly paying off a loan.
Tolerance to the alerting effect builds within about 3–7 days of daily use. It also fades: a week or two of low intake restores most of the response. That makes caffeine unusual among drugs — it rewards being saved.
Elite endurance athletes and shift workers often do exactly this: near-abstinence on ordinary days, a full dose when it counts.
Everyone has a story: espresso is calm, the energy drink is jagged, matcha is "focused," drip coffee makes them anxious. Pick the one that sounds like you and see the candidate explanations.
The molecule is the molecule. FDA notes no difference in how the body handles caffeine that occurs naturally in a plant versus caffeine added to a product. What differs is the dose, the speed, and everything else in the cup.
Three kinds of answer: a real pharmacokinetic interaction (your half-life changes), an additive effect (both drugs push the same direction), or nothing established. Type a drug, or browse.
Not a complete list and not a substitute for a pharmacist. The single most important entry is fluvoxamine: in one controlled study it stretched caffeine's half-life from about 5 hours to about 56.
This is not a "quit caffeine" tool. It's a plan builder. Pick what you want, and it will draft a schedule from the drinks you actually logged above.
Caffeine at sensible doses is a remarkably good drug: cheap, legal, predictable, fast, mildly performance-enhancing, generally well tolerated. The odd part is that we take a real CNS-active drug with unknown dose, unknown timing, daily exposure, tolerance, withdrawal, and no treatment goal. Naming the goal is most of the fix.
Four ways down memory lane. Pick a path; each one is a handful of stops, and they cross at a few points where you can switch tracks.