Quick answer: Probably not, and the popular recipe for doing it is not something the research ever settled on. The internet rule is two parts L-theanine to one part caffeine. In the trials that actually gave people both compounds together, the middle of the range was roughly 82 mg of theanine against roughly 89 mg of caffeine, which is closer to one part to one. Individual trials ranged from about a fifth of a part of theanine per part of caffeine up to about two and a half parts. More importantly, the one trial built specifically to test whether theanine takes the edge off caffeine reported that it did not significantly change jitteriness. What it did change was blood pressure. And the pooled evidence people cite for the stack compares it against a placebo, not against caffeine on its own, which is not the question a coffee drinker is asking.
This is one of those pieces of wellness advice that has travelled a very long way from anything anybody measured. It is worth walking through slowly, because the gap between what the studies did and what the advice says is unusually wide.
The rule everyone repeats, and nobody sources
If you have spent any time in productivity forums, nootropics subreddits or supplement marketing, you have met the ratio. Two hundred milligrams of L-theanine to one hundred milligrams of caffeine. Two to one. It gets passed along with the confidence of a dosing chart, as though it were lifted from a clinical protocol.
It is worth being precise about what that number is and is not. It is a piece of community folklore that stabilised into a rule through repetition. It is not a figure any regulator publishes, and as the next section shows, it is not the ratio the research converged on either. Nobody appears to have decided it. It simply became the thing people say.
That does not automatically make it wrong. Folklore is sometimes right by accident. But it does mean the rule has to earn its keep against the actual trials, and when you line it up against them, it does not survive the comparison.
What the trials actually used
The most complete picture available comes from a 2025 systematic review and meta-analysis in Nutrition Reviews by Payne and colleagues, which searched the literature up to August 2023. Fifty randomised controlled trials met the review's criteria. Sixteen of those trials, reported across fifteen papers, had data usable in at least one pooled analysis, covering 484 participants in total. The median trial enrolled 27 people.
Here is the number that matters for the two-to-one rule. Across the interventions that gave both compounds together, the median theanine dose was 82 mg and the median caffeine dose was 88.74 mg. That is not two to one. That is very close to one to one.
The spread underneath that median is even less flattering to the rule. Drawing from the review's own table of trial characteristics, these are the acute trials that gave both compounds, sorted by how much theanine they used per part of caffeine:
| Trial, as listed in the review | Theanine | Caffeine | Theanine per 1 part caffeine |
|---|---|---|---|
| Yoto 2014, green tea, lower arm | 17.5 mg | 87.47 mg | 0.2 |
| Yoto 2014, green tea, higher arm | 52.2 mg | 191.6 mg | 0.3 |
| De Bruin 2011, black tea | 46 mg | 100 mg | 0.5 |
| Dietz 2017, green tea | 67 mg | 136 mg | 0.5 |
| Baba 2021 | 48 mg | 66.2 mg | 0.7 |
| Dodd 2015 | 50 mg | 75 mg | 0.7 |
| Rogers 2007 | 200 mg | 250 mg | 0.8 |
| Kahathuduwa 2017 and 2018 | 200 mg | 160 mg | 1.3 |
| Haskell 2008 | 250 mg | 150 mg | 1.7 |
| Owen 2008 | 100 mg | 50 mg | 2.0 |
| Einöther 2010 and Giesbrecht 2010 | 97 mg | 40 mg | 2.4 |
Exactly one row on that list is the famous ratio. The rest are scattered across a twelve-fold range, and more of them sit below one to one than above it. The two-to-one rule is not a summary of this evidence. It is one point inside a cloud, promoted to a law.
Two housekeeping notes on the table, because they matter for anyone checking the sources. The review lists the Haskell trial under 2008, which is its print issue; the same paper is indexed to 2007 and appears that way in the reference list at the foot of this article. And these are the acute cognition and mood trials only. The review also covered longer sleep studies dosed daily, which are asking a different question and are not comparable here.
The trial that asked exactly this question
Most research on this combination is really asking "does caffeine plus theanine beat nothing at all?" That is a fine question, but it is not the question a jittery coffee drinker is asking. The question they are asking is narrower: given that I am going to drink the coffee anyway, does adding theanine make it feel less rough?
In 2007, a team led by Peter Rogers at the University of Bristol ran the study designed to answer precisely that. Forty-eight healthy adults were randomised into four groups of twelve, double-blind and placebo-controlled. One group received 250 mg of caffeine, one received 200 mg of theanine, one received both, and one received neither. They rated their own mood and alertness, completed a visual probe task, and had their blood pressure taken before dosing and again starting forty minutes after.
The results, as the authors reported them: caffeine increased self-rated alertness, increased jitteriness, and increased blood pressure. Theanine antagonised caffeine's effect on blood pressure. But theanine did not significantly affect jitteriness, alertness, or other aspects of mood. Theanine also slowed overall reaction time on the visual probe task.
Read that middle sentence again, because it is the whole article in one line. In the trial built to test whether theanine calms caffeine down, the jitteriness did not significantly move. The thing that did move was blood pressure, which is a real and interesting finding, and is not the thing anyone is buying theanine for.
That result deserves the same scrutiny this article has been applying to everything else. Twelve people per group is a small trial, and a null result in a small trial is a weak absence of evidence rather than proof of no effect. A real difference could have been sitting there and gone undetected. It remains the only trial that asked the question directly, against the right comparison, which is why it carries the weight it does here, but it cannot settle the matter on its own. Its funding belongs on the record too, on the same footing as everyone else's: the 2025 review lists this trial as not reporting a funding source, with industry support acknowledged. The funding of the wider set is set out further down, and the one study this article leans on hardest does not get an exemption from the same disclosure.
The authors' own conclusion was appropriately careful. They wrote that theanine is a physiologically and behaviourally active compound, that it is unclear how its effects might explain the perceived difference between tea and coffee, and that the evidence suggests it may be useful for reducing raised blood pressure. Notice what is absent from that summary.
That last part needs bounding, because it is easy to misread. It is a suggestion about where the research might go next, not a finding about anyone's medical care. Their trial measured a single dose in healthy adults over about forty minutes. It was not a study of managing blood pressure, and nothing in it is a reason to change anything a clinician has advised.
The other 2007 trial, where the combination helped and theanine alone hurt
The same year, a group at Northumbria University led by Crystal Haskell ran a balanced crossover trial with twenty-four participants, giving 250 mg of theanine and 150 mg of caffeine alone and in combination, with salivary caffeine monitored to confirm dosing.
This trial is genuinely good news for the combination on task performance. Caffeine alone produced faster digit vigilance reaction times, better accuracy on a demanding visual processing task, and blunted the rise in self-reported mental fatigue. The combination did all of that and more: faster simple reaction time, faster numeric working memory reaction time, better sentence verification accuracy, and a significant interaction between the two compounds on delayed word recognition. Headache and tiredness ratings went down and alertness ratings went up.
Now the part that usually gets left out of the summaries. Theanine on its own increased headache ratings, and it decreased the number of correct answers on a serial sevens subtraction task. On its own, in this trial, it made a mental arithmetic task worse and headaches slightly more likely.
That is not a reason to be frightened of theanine, which has a broadly unremarkable safety record at the doses used in research. It is a reason to be suspicious of the way it is usually described. A compound routinely marketed as pure downside-free calm produced two unflattering results in one of the most-cited trials about it, and those results are rarely mentioned.
What the 2025 meta-analysis pooled, and the comparison it never made
When Payne and colleagues pooled what could be pooled, the combination did beat placebo on several measures in the first two hours. Attention switching accuracy in the second hour came out at a standardised mean difference of 0.33, with a confidence interval from 0.13 to 0.54. Digit vigilance accuracy in the second hour was 0.20, from 0.02 to 0.38. Choice reaction time in the first hour was −0.48, but its interval ran from −1.01 to 0.05, which crosses zero. Overall mood in the second hour was 0.26, from −0.10 to 0.63, which also crosses zero.
Two of the pooled trials reported jitteriness directly, measured on a visual analogue scale two hours after dosing. The combination came out 3.25 units lower than placebo, with an interval from −3.81 to −2.68. Headache came out 6.83 units lower.
Here is where you have to read carefully, because this is the result most likely to be quoted at you as proof the stack works. That jitteriness figure compares theanine plus caffeine against placebo. It does not compare theanine plus caffeine against caffeine on its own.
Those are completely different questions. "Is a caffeine-and-theanine capsule less jittery than a sugar pill?" tells you nothing useful about whether adding theanine to your coffee helps, because the coffee is the caffeine. To answer that you need the combination measured against caffeine alone, and the review's pooled jitteriness analysis does not report it.
The review also notes that no minimal important difference appears to have been established for any of these outcomes. In ordinary language: nobody has worked out how big a shift on that scale a person would actually notice. A statistically detectable three-unit difference on a hundred-unit line may or may not be something you could feel. That is not a criticism of the researchers, who were right to say so. It is a limit on how much weight the number can carry.
Calmer than nothing is not the same as calmer than coffee
It is worth spelling out the logic, because it is the single most common error in this whole topic and it is easy to make.
Imagine you want to know whether adding a passenger makes a car quieter. You could compare the car-with-passenger against an empty driveway. The driveway is quieter, or the car is louder, and either way you have learned nothing about the passenger. The only comparison that answers your question is car-with-passenger against car-alone.
Almost all of the pooled evidence on this combination is the first kind of comparison. It is measured against placebo, because that is how you establish that a compound does anything at all. It is the right first study to run. It is simply not the study that answers the jitters question, and reading it as though it were is where the popular version of this advice goes wrong.
The one study that did run the right comparison, on the right outcome, was Rogers 2007. It found no significant effect on jitteriness.
How strong is any of this, really
Not very, and the review is admirably direct about it. Fifty trials were screened and only sixteen could be pooled at all. Only three of the pooled analyses included more than four studies. The jitteriness result rests on two. Across the whole set, 484 people were studied, with a median of 27 per trial.
The review describes its own overarching limitation as an inability to draw definitive conclusions about the true effects of either the combination or theanine alone, pointing to the large prediction intervals. Its heterogeneity estimates frequently ran from a lower bound of zero to an upper bound above 75 percent, meaning it was often unclear how much of the variation between studies was real and how much was noise.
Funding is worth stating too. Among the pooled trials, industry was the most common funding source at eight, followed by academia at four and government at two, with four studies not reporting their funding at all. Two of the review's own authors were, per the paper's affiliations, employed by tea companies. None of that makes the findings wrong, and the review is transparent about all of it, which is to its credit. It is context a reader is entitled to have.
One more limitation that rarely gets mentioned. Almost none of this was tested as a drink. The review observes that few studies used tea or tea-equivalent doses and ratios, that no study in the relevant analyses delivered its intervention as tea or a tea-flavoured drink, and it explicitly calls for research in free-living contexts rather than laboratories. The evidence base for a beverage habit is built mostly on capsules taken in a lab.
One line from this review gets quoted more than any other, so it is worth saying where it sits. The authors open by noting that the evidence available to them pointed toward the benefits of tea, or of theanine plus caffeine, being mostly driven by caffeine. That is the premise their review set out to test, not a result it produced, and their own conclusion is more generous: compared with placebo, the combination likely improves performance on attention tasks and may benefit alertness and overall mood, and theanine on its own may improve performance on the choice reaction and rapid visual information processing tasks. Where the caffeine signal does show up is in the earlier 2014 meta-analysis by Camfield and colleagues, whose moderator analysis showed trends toward greater change in effect size for caffeine dose than for theanine dose, particularly in the first hour.
So, does theanine cancel coffee jitters?
On the evidence available: no, not in the way the advice promises, and the ratio attached to that advice is not the one the trials used.
The honest version is narrower and less exciting. Caffeine and theanine taken together outperform a placebo on some attention measures in the first couple of hours, on small numbers of small trials with wide uncertainty. In the Rogers trial, theanine blunted caffeine's blood-pressure rise. The jitteriness question, asked directly and properly controlled, came back not significant.
If someone tells you the stack reliably removes the edge from coffee, they are describing a study that has not been published.
What actually changes how a coffee feels
Here is a thought that follows directly from the studies above, rather than from anything sold in a bottle.
Look at what those trials had to control in order to measure anything. Fixed single doses, weighed out. Caffeine washout periods before testing, running to twelve hours or more in several of the trials. Quiet laboratory rooms. Capsules rather than drinks, precisely so that taste, temperature, volume and drinking speed could not interfere.
Every one of those controls exists because the researchers knew those variables move the outcome. They removed them to see the compound clearly. Your Tuesday morning has none of that control, which means the variables the trials carefully eliminated are exactly the ones still operating on you, and unlike the compounds, they are free.
The size of the serving is the first one. It is the variable the evidence above keeps pointing back at: the 2014 moderator analysis showed effect sizes trending more with caffeine dose than with theanine dose, and caffeine is the compound that raised jitteriness in the Rogers trial. A larger serving is a larger dose of the thing producing the jitteriness in the first place. That is not advice about a number; it is just the direction the arrow points.
Speed is the second. A cup drunk in four minutes and the same cup drunk across an hour deliver the same amount to a very different curve. The trials gave a single bolus dose because a slow, variable intake would have blurred their measurements. Outside the lab, that blurring is available to you.
Food is the third. Every one of these trials tested on a controlled stomach for the same reason. A coffee on an empty stomach first thing and the same coffee alongside breakfast are not the same event.
Habituation is the fourth, and it is the one people forget. Several of these trials required participants to go without caffeine for twelve hours or more before testing. That is a deliberate and necessary design choice, but it means the participants were, to some degree, caffeine-deprived when measured. If you drink coffee every day, you are not that person, and you should expect your own experience to differ from theirs. We have written separately about what happens when coffee stops seeming to work, and about how much caffeine is generally considered too much.
Timing is the fifth. When in the day a serving lands changes both how it feels and what it does to the night that follows. That is its own topic, and we have covered it in a guide to when to stop drinking coffee, a look at the afternoon energy dip, and a piece on matching coffee to your own daily rhythm.
None of that is as satisfying as a ratio. It is, however, supported by the same trials that fail to support the ratio, which seems like the better trade.
If you enjoy watching a confident wellness rule come apart under its own sources, the ninety-minute cortisol rule is a close cousin of this one. For background on the compound doing most of the work here, see our explainer on what caffeine actually is, and for the wider habit of combining coffee with other botanicals, our guide to stacking. If you have ever blamed a bad night on an afternoon cup, our piece on what sleep trackers can and cannot see is worth a read.
Where CafeBank fits, and where it does not
Plainly: this article is not a route to anything we sell. CafeBank makes coffee. We do not sell L-theanine, we do not add it to our coffee, and nothing above should be read as being about a CafeBank product.
We wrote it because a large number of people are being told to buy a second product to fix a problem with the first one, on the strength of a ratio that the research does not support. Saying so costs us nothing and seems more useful than staying quiet about it.
Who should take extra care
Individual responses to caffeine vary a great deal, and some of that variation is not something you can train your way out of. If a normal serving reliably leaves you feeling unwell, that is worth taking seriously as information rather than as a problem to be solved with an additional supplement.
People with poorly controlled hypertension, atrial arrhythmias, or diagnosed anxiety disorders should consult their clinician before adding any caffeine source.
These statements have not been evaluated by the Food and Drug Administration. Nothing in this article is intended to diagnose, treat, cure, or prevent any disease. This article is general information about food ingredients, not medical advice. If you are pregnant or nursing, take prescription medication, or have a heart, thyroid or anxiety condition, talk to your doctor before adding a caffeinated botanical product to your routine.
Frequently asked questions
Is the 2:1 L-theanine to caffeine ratio backed by research?
No. In the trials that gave both compounds together, the median doses were about 82 mg of theanine to about 88.74 mg of caffeine, which is close to one to one. Individual trials ranged from roughly 0.2 to 2.4 parts theanine per part caffeine. Only one trial in that set used exactly two to one.
Did any study test whether theanine reduces caffeine jitters directly?
Yes. A 2007 randomised, double-blind, placebo-controlled trial of 48 adults by Rogers and colleagues gave 250 mg of caffeine, 200 mg of theanine, both, or neither. Caffeine increased jitteriness. Theanine did not significantly affect jitteriness, alertness or other aspects of mood.
Does theanine do anything at all, then?
It does. In that same trial it antagonised caffeine's effect on blood pressure. In pooled analysis, theanine alone was associated with faster choice reaction time in the first hour compared with placebo. It also slowed reaction time on a visual probe task in one trial, and in another it increased headache ratings and reduced correct answers on a serial sevens subtraction task. It is an active compound with mixed effects, not an inert one.
Why do so many sources say the combination works?
Because it does beat placebo on several attention measures in the first two hours, and that is what most of the studies measured. The gap is between "better than nothing" and "makes your coffee feel smoother", which requires comparing the combination against caffeine alone. That comparison is largely missing for jitteriness.
How much theanine is in a normal cup of tea?
The 2025 review notes that 50 mg of theanine, the lowest dose used in any trial in its relevant analyses, is roughly equivalent to two cups of black tea. Most trials used more than 100 mg, and none of them delivered the dose as tea.
Is theanine dangerous?
Nothing in the trials discussed here suggests it is, at the doses studied, in healthy adults. The unflattering findings above are about effectiveness and minor side effects, not safety. Anyone taking medication or managing a health condition should raise it with their own clinician rather than rely on a general article.
Does coffee contain any theanine?
No. Theanine is found in tea, from the Camellia sinensis plant. It is not a coffee compound, which is why the advice always involves adding it as a separate supplement.
What is the single most useful thing to take from all this?
That the amount of caffeine, and the speed and circumstances in which you drink it, are doing far more work than any ratio. The 2014 meta-analysis showed effect sizes trending more with caffeine dose than with theanine dose, and in the Rogers trial it was caffeine that raised jitteriness.
References
1. Payne ER, Aceves-Martins M, Dubost J, Greyling A, de Roos B. (2025). Effects of Tea (Camellia sinensis) or its Bioactive Compounds l-Theanine or l-Theanine plus Caffeine on Cognition, Sleep, and Mood in Healthy Participants: A Systematic Review and Meta-Analysis of Randomized Controlled Trials. Nutrition reviews. 83(10):1873-1891. DOI: 10.1093/nutrit/nuaf054. PMID: 40314930.
2. Rogers PJ, Smith JE, Heatherley SV, Pleydell-Pearce CW. (2007). Time for tea: mood, blood pressure and cognitive performance effects of caffeine and theanine administered alone and together. Psychopharmacology. 195(4):569-77. DOI: 10.1007/s00213-007-0938-1. PMID: 17891480.
3. Haskell CF, Kennedy DO, Milne AL, Wesnes KA, Scholey AB. (2007). The effects of L-theanine, caffeine and their combination on cognition and mood. Biological psychology. 77(2):113-22. DOI: 10.1016/j.biopsycho.2007.09.008. PMID: 18006208.
4. Camfield DA, Stough C, Farrimond J, Scholey AB. (2014). Acute effects of tea constituents L-theanine, caffeine, and epigallocatechin gallate on cognitive function and mood: a systematic review and meta-analysis. Nutrition reviews. 72(8):507-22. DOI: 10.1111/nure.12120. PMID: 24946991.