That dry, puckering feeling a cup of coffee leaves behind is not a taste at all. It is touch. What you are feeling is the thin film of protein that keeps the inside of your mouth slippery being stripped out from under the surfaces it was lubricating, so those surfaces start to drag against each other. Bitterness happens at a receptor on your tongue. This does not, which is why it keeps going after you swallow and why a sweeter coffee does not fix it.
The compounds responsible in coffee are largely built during the roast, out of the wreckage of chlorogenic acid. That is a different origin story from the one a drinker usually gets, and it changes which knobs are worth turning.
Is astringency a taste?
No, and there is a clean anatomical reason. You feel it on your soft palate, on your gums and on the inside of your lips. There are no taste buds in any of those places.
Alessandra Rinaldi and Luigi Moio put it plainly in their 2020 review of salivary protein and tannin binding: astringency "can be defined as a tactile sensation, because: (i) it is perceived on non-gustatory surfaces such as on the soft palate, gingiva, lips." If a sensation shows up where the taste system does not reach, it is not the taste system reporting it.
Dr Ramya Mohan, a quality specialist at the Coffee Board of India, gave the same thing its proper name in New Ground Magazine, describing "a harsh, tactile sensation that decreases salivary lubrication and causes a papery dryness on the tongue." Trigeminal, not gustatory. Your fifth cranial nerve carries touch, pressure and temperature from your face and mouth, and it is the nerve doing the reporting here.
This matters at the bar more than it sounds like it should. If you taste a cup as harsh and call it bitter, you will reach for the remedies that work on bitterness. Sweetness, a shorter brew, a softer roast. Some of those help a little, for reasons I will get to. None of them is aimed at what is actually happening.
What is actually happening in your mouth?
Saliva is not just water. A large share of what is in it is a family of proteins built around the amino acid proline, and their job is lubrication. Rinaldi and Moio note that proline-rich proteins "account for approximately 70% of the total secretory protein." Proline is what they call an alpha-helix breaker, which leaves the protein flopping about in a loose random coil rather than folding up tight, and that loose shape is exactly what lets it grab hold of almost any polyphenol that comes past.
When a polyphenol-heavy liquid arrives, those proteins bind it, clump, and drop out of solution. The lubricating film goes with them. A 2026 paper in Scientific Reports by Ammam and colleagues measured the mechanics of it directly and found that the resulting complexes cause "a loss of lubrication, which leads to an increase in friction forces that activate the mechanoreceptors." Mechanoreceptors. Touch sensors. The same class of nerve ending that tells you a shirt label is scratching your neck.
The part of that paper worth carrying around is what the proline-rich proteins are for. They are not the casualty. They are the defense. The authors describe them binding tannins "before they can aggregate with salivary mucin proteins, thereby preserving lubrication" — a sacrificial protein whose whole purpose is to be ruined first so the mucins that actually do the lubricating are spared. Your mouth budgeted for this.
Which is why the feeling builds across a session rather than arriving all at once. The third cup of a harsh coffee is worse than the first because the sacrificial layer has already been spent on the first two.
Which compounds in coffee cause it?
Scott Rao, who has written more useful pages about brewing than almost anyone, names the class: "Large molecules called polyphenols, in particular chlorogenic acids (CGAs) and tannins, are the likely, primary sources of astringency in brewed coffee." Jonathan Gagné, writing at Coffee ad Astra, describes the binding step the same way and adds the size argument: polyphenols "are mostly very large molecules, which makes them heavier and harder to extract from coffee particles, compared to smaller molecules like caffeine and the various acids we enjoy."
Both are right, and there is a step between them that changes the picture. Chlorogenic acid sits in green coffee at 5 to 8 percent of dry weight, a figure we have published twice before, in our piece on what makes roasted coffee brown and again in the one on coffee chaff. It does not survive the roast. It comes apart, and what it comes apart into gets built into something else.
In 2025 a team at Ohio State, led by Linne, Tello, Simons and Peterson, isolated the astringent fraction out of brewed coffee and characterized it. Writing in Food & Function, they describe "a complex, phenolic-dominant, melanoidin fraction" whose components "result from derivatization of chlorogenic acid thermal reaction products during roasting." A trained panel could pick the isolate out of plain water at 7.9 milligrams per liter, and could still pick it out when it was spiked back into coffee.
Read that slowly, because it reverses the usual order. The astringent agents in your cup are not chlorogenic acid sitting in the bean waiting to be extracted. They are melanoidins, the big brown polymers roasting builds, assembled partly out of what chlorogenic acid turned into on the way. Green coffee brings the raw material. Roasting makes the thing you feel.
One caution, because our own writing has to stay straight on it. Our page on malic acid and our cultivar notes both treat chlorogenic acids as something that reads in the cup as brightness and then as bitterness, and that is still true. Bitterness and the dry feeling are two different reports from two different systems, and one family of compounds can feed both. Saying a coffee is astringent is not another way of saying it is bitter, and a cup can arrive with plenty of one and almost none of the other.
Does a darker roast make coffee more astringent or less?
I do not know — and I have not found anyone who has measured it properly.
Here is the state of the evidence. Chlorogenic acid falls steadily as roast advances, which a 2024 study in Foods by Kim and colleagues tracked against roast color, watching the acid decompose and its products accumulate. Sensory work by Birke Rune and colleagues in Current Research in Food Science found that for coffee's acids, "the astringency and mouthfeel of the different organic and inorganic acids were more important in the description of the acids than the bitterness and sourness." Those acids are being felt more than they are being tasted. And the Ohio State work shows the astringent melanoidins accumulating as roasting proceeds.
So one input goes down while its products go up, and nobody I can find has run a roast ladder against a trained panel to say which effect wins. Rao attributes it to underdevelopment in the roast. Others put it at the brewer. They cannot both be the whole story, and neither of them cites a measurement.
We grade roast color with an Agtron on every batch as a standing practice, which is how we keep one variable from wandering while we argue about the others. It also has a published limit we say out loud on our toll roasting page: the green decides what the roast cannot. A new crop year or a different lot tastes different no matter how steady the color reads. Holding the roast still tells you the dryness in your cup came from somewhere else. It does not tell you where.
Why does a quaker taste dry?
This is the one place our own writing has already touched it. Our article on what a quaker is says a quaker brings "lower sweetness, a thinner body, a dry astringent finish and a stale peanut note." That line is accurate and it has been sitting there since the first of September without explaining itself.
Here is the explanation. A quaker is a seed from a cherry picked before it ripened, so it never built the sugars and never developed the precursors that brown properly in the roast. It stays pale while everything around it darkens. What it does carry is the defensive phenolic chemistry a green seed is full of, and what it lacks is the sweetness that would otherwise mask the drag those compounds cause. Sweetness is genuinely protective here, and Barista Hustle is right that it reduces the perception, so a seed that is short on sugar and long on phenolics is the worst possible combination for this specific sensation.
Specialty grade permits zero quakers, as we have published: the FAO's green classification allows none in Specialty and three in Premium, with no primary defects and not more than five full defects in a 300-gram sample. Those are figures per graded sample rather than per retail bag, and one pale bean in a batch is hard to notice. A handful of them will dry out a whole pot.
What can you change at the brewer?
Grind, dose, and how evenly the water gets through the bed — in roughly that order.
Gagné's argument here is the most useful thing published on the subject and I will hand it to him rather than restate it as ours. In his 2022 follow-up he concludes that these compounds are poorly soluble, that they "float around freely in the slurry, and avoiding astringency in the cup is a matter of filtration", and that "grinding too fine will liberate a lot more astringent compounds." He is describing something closer to a plumbing problem than a chemistry problem. The compounds are there either way; the question is whether the bed traps them or lets them through.
That squares with what we already tell people about grind. Our article on whether a finer grind makes coffee stronger puts it this way: when a bed packs too tight, water finds the loosest path, and "some grounds over-extract and taste bitter and drying while others are barely touched." The drying half of that sentence is the subject of this article. Same defect, same fix.
Dose is the one people skip. A cup that tastes harsh is very often a cup that was under-dosed and then over-run to make up the volume, and the gap between a careless measure and a careful one is larger than it looks. We publish a number for that on our restaurant page: a level scoop and a heaped scoop differ by 40 percent. If the dose moves by that much, no grind setting is going to rescue it. Weigh it, then adjust, which is the same order we recommend for brew ratios generally.
Water temperature and contact time both move it as well, and both move it less than the first two. Start with the ones that do the work.
Why doesn't our own label tell you any of this?
Because our label publishes flavor and never texture, and I think that is a real gap rather than a considered choice.
Every coffee we sell carries tasting notes, and every one of those notes is a flavor. Brown sugar. Dried tamarind. Cranberry. Clove. Not one coffee on our shelf carries a single word about how the cup feels, and I checked the whole catalog rather than guessing. A cup score does not cover it either. It is a quality judgment made under a protocol, and it collapses everything a coffee does into a number that cannot tell you whether the finish will dry you out.
So a drinker who is sensitive to this gets nothing from us at the point of purchase. Somebody who finds most light roasts hard work in the mouth, or who wants to know whether a coffee finishes clean or finishes rough, has to buy the bag and find out. Roasters in general are not good at this, us included. I have also seen us call a finish "refreshingly tannic" on one of our own coffees and mean it as a compliment, which it was. The drag can read as structure when there is enough sweetness under it, the same way it does in tea or in a young red wine. But the word is doing two opposite jobs in our own copy, and we have never once told anybody which one we meant.
Questions we get about astringency
Is astringency the same as bitterness?
No. Bitterness is a taste, registered by receptors on your tongue. The dry, puckering feeling is touch, registered by mechanoreceptors when polyphenols strip the protein film that lubricates your mouth. They often turn up together because the same brewing fault can produce both, but they are separate sensations and you can get either one without the other.
Does over-extraction cause it?
Uneven extraction is the usual culprit rather than high extraction as such. When a bed is ground too fine or packed unevenly, water channels through the loose parts and hammers the rest, and the compounds responsible get liberated and carried through instead of being trapped by the bed. Coarsening the grind and pouring so the bed stays level fixes more cups than shortening the brew does.
Why does the feeling get worse as I keep drinking?
Your saliva carries proline-rich proteins that bind these compounds before they can reach the proteins that actually lubricate your mouth. That defense is consumable within a sitting. Once it has been spent on the first cup, the second and third arrive with less protection, so the same coffee feels rougher as you go.
Can a light roast be astringent?
Yes, and a dark one can be too. The astringent compounds in coffee appear to be assembled during roasting from chlorogenic acid's breakdown products, so both the raw material and the product shift as roast advances. No published study we have read has run a proper roast ladder against a trained panel, so anyone giving you a confident answer about roast level is ahead of the evidence.
Does a quaker make coffee astringent?
A quaker contributes a dry finish, yes. An unripe seed never built the sugars that would soften the sensation while keeping the defensive phenolics that cause it. One is hard to notice. A handful flattens the cup and dries it out.
Where to get a coffee ground for your brewer
If your cup has been drying you out, the cheapest thing you can change is the grind, and we will do it for you. Tell us what you brew on and we grind to it at no charge, at roast time, so the beans are not sitting ground while they wait for you. ASCENT, our house blend, cups at 92.3 and is what we point most people at first.
