A cloth filter is the only part of a brewing kit that can spoil while nobody is touching it. Leave one to dry on the edge of the sink over a weekend and the oils held in it go stale, and the next several cups arrive with that staleness sitting on top of them before the coffee gets a word in. That is the standing cost of the method, it is owed daily, and it is the decision you are actually making. The fibre question comes second and it is the smaller one.
Buy brushed cotton, usually sold as flannel or as a nel filter, if you want the heaviest and roundest cup and you accept that you are replacing it every few months. Buy plain-woven linen if you want quicker drainage, almost no lint, and a filter whose behaviour drifts more slowly over its life. The fibre name on the label matters less than the nap: a brushed cloth filters through its whole thickness, a smooth cloth filters mostly at its face, and that single structural difference moves the body of the coffee further than cotton and flax do.
Both hold fine particles back and let oil through. That is the entire reason cloth exists as a category, and neither fibre changes it.
Body is a touch sensation, and it is made of what water cannot dissolve
Before the weave can matter, it helps to be precise about the thing it is supposed to change, because "body" gets used for four different sensations in the same conversation.
Jessica Easto, in How to Taste Coffee, follows the Specialty Coffee Association's Coffee Sensory and Cupping Handbook and keeps body narrow: it is thickness plus texture, and it sits inside the wider category of mouthfeel, which also covers temperature and astringency. Thickness is the weight the liquid seems to carry as it moves across the tongue. Texture is the qualitative half, whether that weight arrives as silky, gritty, drying or oily. Both are touch. Neither is taste, and the nerve doing the work is not the one carrying flavour.
What produces them is everything in the cup that the water never took into solution. Easto's account, which the SCA handbook underwrites, sorts that insoluble material into three classes, and a filter treats each of them differently.
The bulkiest are the polysaccharides: cellulose, hemicellulose, arabinogalactan, pectin. These molecules are simply too big to go into solution, so they ride in the water as loose suspended structures instead of dissolving into it. Then there are the fines, the smallest debris left by the grinder, which stay aloft while the cup is in motion and drop to the bottom once it is still. Last are the lipids, coffee's oils, which repel water and so never dissolve at all. They stay dispersed through the brew only because other compounds in it hold them there. That last class is what separates one filter medium from another, because oil is the thing paper takes out and cloth lets past.
None of that is dissolved. The dissolved fraction, the part that sets strength and carries most of what you would call flavour, passes through every filter you can buy. So a filter is not arguing with you about flavour. It is arguing about what is floating in it.
Then comes the part that makes this worth an article rather than a footnote. Thickness does not sit politely beside taste. Easto reports, from the sensory research she cites, that a thicker liquid needs more sour, more sweet or more bitter in it before the tongue registers any of the three at all, so each of them reads weaker at a concentration that has not changed. Umami is the exception and gets louder. The effect runs in the other direction too: a sweeter cup is judged thicker than it is, a more acidic one thinner. A heavier cup therefore reads less sharp while holding exactly the same acid at exactly the same strength. You have not changed the chemistry of the drink. You have changed how much of it your mouth reports.
That is why people who move to cloth describe the coffee as rounder and often as sweeter, and why the description survives blind. It is also why a cloth filter is a poor tool for fixing a cup that is sour because of how it was extracted rather than because of what it passed through.
What a weave is doing, mechanically
Cloth catches particles two ways, and the difference is the whole argument.
Surface filtration is a sieve. The openings between yarns decide what passes, and a particle either fits or does not.
Depth filtration is a maze. Particles travel a long, twisting path between fibres and are caught by contact with fibre surfaces rather than by size alone. Depth filters catch particles far smaller than their apparent holes, they hold much more material before they block, and they run slower.
Cotton is a seed hair, a short staple fibre roughly twenty to thirty-five millimetres long in the common upland varieties, nearly pure cellulose, flat and twisted along its length with a hollow core. Flannel is cotton that has been mechanically brushed after weaving so those short fibre ends are pulled up into a raised nap. That nap is not decoration. It converts a woven fabric into a depth filter with a real thickness to it.
Flax, the plant linen comes from, is a bast fibre taken from the stem rather than the seed. Its fibres are longer, straighter, stiffer and smoother-surfaced, bonded into bundles that run much further than a cotton staple does. Plain-woven linen has very little to brush up, so it behaves closer to a surface filter, with a modest depth component from the yarn itself.
Four consequences follow, and they are the ones you will actually notice.
Flow. A napped cotton drains more slowly than a comparable linen at the same grind and dose. If your drawdown is already slow, cotton will make it slower, and you will need to grind coarser to keep the brew inside the three to four minutes Anette Moldvaer gives for a cloth drawdown in Coffee Obsession, adjusting grind and dose until the time comes right.
What gets through. Depth filtration holds back more fines, so a brushed cotton produces a cup with body and comparatively little grit. Linen passes more of the small stuff, which reads as fuller and leaves slightly more sediment in the last mouthful.
Oil loading. The nap has vastly more internal surface than a smooth weave, and coffee oil clings to it. A new flannel filter is at its cleanest and fastest on day one, tends to be at its best after a couple of weeks of use, and then slows as it loads. Linen loads too, more slowly, and from a lower ceiling.
Ageing. Cotton nap mats down with heat, handling and repeated wetting, so the filter you bought is measurably not the filter you own in month three. Linen holds its structure longer because the fibre bundles are longer and stiffer to begin with. This is the strongest practical argument for linen and it has nothing to do with flavour.
What we have not measured
Everything in the section above is mechanism, drawn from how filtration works and from what these two fibres are. It is not measurement, and we are not going to dress it up as one.
Here is the test that would settle it, written down so that we can be held to it. Same roast, same grind setting, same dose, same water and the same pouring pattern, brewed through a brushed cotton and a plain linen filter on the same frame, alternating order across at least six pairs. Record drawdown time to the second, read total dissolved solids with a refractometer, and run the pairs blind past people who did not brew them.
Our prediction, stated in advance so it can be wrong: dissolved solids land within a small margin of each other, drawdown differs by tens of seconds, and tasters separate the pairs on texture at a rate well above chance while separating them on strength at close to chance. If the numbers come back the other way, the correction goes at the top of this article rather than at the bottom.
Until then, take the fibre comparison as reasoning, and take the care protocol below as documented practice, because Moldvaer and Racineux and Tran both set out the same instructions independently.
Fibre by fibre
| Cloth | Body it gives | Flow | Lint | Useful life | Buy it if |
|---|---|---|---|---|---|
| Brushed cotton (flannel, nel) | Heaviest and roundest, very few fines | Slowest | Sheds, especially when new | Shortest. Nap mats and loads with oil | You want the fullest cup cloth can give and you brew most days |
| Plain-woven cotton | Fuller than paper, some fines through | Moderate | Some | Moderate | You want part of the effect with less maintenance |
| Linen | Full, with slightly more sediment | Fastest | Almost none | Longest | You are inconsistent, or you want the same behaviour in month six |
| Cotton and linen blends | Between the two, depending on the mix | Between | Low to some | Between | Untested by us, and the label rarely states the ratio |
One argument you will meet and we cannot settle: whether a napped filter should be used with the fuzzy side facing the coffee or facing out. Experienced nel drinkers hold both positions with conviction. We have no measurement, so we have no opinion, and neither does anyone who has not run the pairs.
The care protocol is a purchase criterion
This is the part that decides whether a cloth filter is a good buy for a particular person, and it is documented rather than argued.
Rinse a new filter before its first use. Wash it in hot water only and never with soap, which is Moldvaer's instruction and ours: the residue is not fully rinsable and the cloth carries it into the coffee. Store it wet between uses, sealed in the fridge or the freezer, and never leave it to dry out on a draining board. A cloth that has dried out picks up a stale, musty taint, and it transfers to every brew afterwards.
That protocol is old. William Ukers recorded the same instruction in 1922, when cloth was simply what filters were made of and the trade kept its urn bags submerged in cold water rather than letting them dry. The paper filter's real achievement, when Melitta Bentz patented hers in 1908, was disposing of a wet souring textile rather than improving the cup. If you buy cloth, you are opting back into the problem paper was invented to solve, knowingly, for the texture.
The fibre choice interacts with this directly. A napped cotton holds more water and more oil, so it goes off faster when you skip a day. Linen is more forgiving of an owner who is not reliable. If you already know you are the kind of person who leaves things by the sink, that is a better reason to buy linen than any claim about body, and it is the recommendation we would make to most people who ask. The full protocol, including what to do when you have already ruined one, is in the care piece on why a cloth filter must never dry out.
What this will not do
It will not fix extraction. A sour cup through linen is a sour cup with better texture, and a bitter one is bitter with better texture. Grind and contact time set balance, and no weave overrides them.
It will not make the coffee stronger. Same dose and same water give roughly the same dissolved solids whatever the cloth. What changes is thickness, and thickness reads as intensity even though it is not. This is one of the four unrelated things people mean by the word strong, and cloth only touches one of them.
It will not fail cleanly. A filter that is going off does not stop working. It drifts, over weeks, and because it drifts slowly you adapt to it and blame the beans. Mark the date you started using it somewhere you will see it, because your palate will not tell you.
It will not give you what paper gives you. If what you like about a paper-filtered cup is separation, the sense of being able to name the parts, cloth moves you away from that and not towards it. The comparison across all three filter media is worth reading before you spend anything, because a decent number of people who buy cloth turn out to have preferred paper all along.
And it will not settle anything about what an oilier cup does to a person. Cafestol and kahweol are two lipids carried in coffee oil, and Urgert and Katan's quantitative review of human intervention trials, published in the Journal of the Royal Society of Medicine in 1996, established both that they raise serum cholesterol and that a paper filter physically retains them. The effect tracks how much unfiltered coffee somebody drinks in a day, and it reverses once the coffee is withdrawn. What nobody appears to have published, as of August 2026, is a measurement of how much of that oil a cloth filter holds back. Cloth sits between paper and metal in the brewing manuals, and that is a description of how the cup tastes rather than a measurement of the compounds in it. So we do not know where cloth lands on that scale, and we are not going to guess it for you. If you have been advised to manage LDL cholesterol, that conversation belongs with the clinician who has your history in front of them, and this article is information about filters rather than medical advice.
Two cloths, two purchases
The word cloth covers two different objects and shops rarely separate them.
A brewing filter is a shaped, seamed, usually napped piece sized to a frame, and it lives under the wet-storage protocol above. A straining cloth is a flat rectangle used once per batch and then washed, most often to strain immersion cold brew through a sieve, which is exactly what Racineux and Tran specify for that method. A straining cloth wants an open plain weave and no nap at all, because a nap will trap a punishing amount of oil in one contact with a whole batch of cold concentrate.
Do not use your brewing filter to strain cold brew. One session will load it more than a month of ordinary use.
The station
Cloth is the one filter medium that only works inside a complete arrangement, because the storage container is not optional.
| Part | Why it is here | Ours |
|---|---|---|
| Burr grinder holding a medium setting | Cloth is a medium grind, and drawdown is your only feedback signal | No |
| Cloth filter and its frame | The decision this article is about | No |
| Scale | Cloth runs at roughly sixty grams of coffee per litre of water, the same number as every other filter method in the house | No |
| Small sealed jar, kept in the fridge | Where the wet filter lives between brews. Without it the method fails inside a fortnight | No |
| A cup you want to hold | Cloth coffee is oilier and reads darker, so it shows nothing in glass and gains from a vessel that holds heat | Yes |
Five things, and we sell one of them. We do not sell filters, frames, grinders or scales, and this piece would read differently if we did.
The last part is the only one where we have a stake and it is worth one honest sentence. A cloth-filtered coffee is opaque, so the transparency argument for glass does not apply, and the cup's material and wall thickness are doing more work here than with a lighter brew.
Buy the cheapest respectable linen filter first, keep it alive for a month, and only then decide whether you want the heavier cotton version. If you want to know which you actually prefer rather than which you expect to prefer, taste them against each other in the same session rather than a week apart. Memory for texture is short, and the second cup on a different morning is not a comparison.