coffee/honey-process · 14 min

coffee/ · 14 min

What honey process means

Not a third method. A dial, and a decision about how much of the fruit to leave on.

coffee/honey-process -> main94d4e59 · 24 Sep 202629 sourcesask your agent about this post

00 read this with your agent

Read this with your agent first.

Copy one of these, paste it into whatever AI you use, and it will teach you from this page and only this page, at your level. An agent in a browser that supports WebMCP can use the page’s tools directly.

Teach me from scratch

What does 'red honey' on a bag mean?

01 the middle route

There are not three ways to process coffee. There is one dial.

Honey process is a decision about how much of the fruit stays on the seed while it dries.

A coffee bean is a seed inside a small fruit that looks like a cherry. Working inward there is a skin, a layer of sweet pulp, a thin coat of slippery sugar called mucilage that clings to the seed like the flesh on a mango stone, a papery shell called parchment, and then the seed. The full anatomy belongs to the post on the dry route; three sentences is all this one needs.

Processing gets you from that fruit to a dry seed that keeps, and every route does the same two things in a different order: remove the fruit, and remove the water. The washed route takes the fruit off first, in a machine and then in water; that is the wet route. The natural route takes nothing off and dries the whole cherry.

Honey process is the route in between, and it is defined by what it does not do.²⁸ The cherry goes through a pulper, the machine that squeezes off the skin and most of the pulp, exactly as a washed coffee does. Then the washing is skipped, and the seed goes out to dry with the mucilage still stuck to it. The sugar a washed mill rinses down a drain stays on the bean.

So honey process is not one thing. A mill can set its pulper to scrape off nearly all the mucilage, or hardly any, or anywhere in between, and each setting gives a different coffee. That is the dial. White, yellow, red and black are attempts to put marks on it, and the rest of this post is partly about how badly those marks are drawn.

fig 1The whole post in one axis. Scrub from a washed bean to a whole cherry and the readout names the grade, the drying time and what the cup does at that setting. Only the ends have days: sun-dried washed parchment takes 8 to 10 days and a natural up to 4 weeks ²⁹. The honey grades fall between them in order only, and black can take about twice as long as white ²². The grade positions are my own synthesis of three guides that do not agree; the next figure shows how far apart they sit. Hover, drag or arrow along the axis.FAO ends, trade middle · interactive

02 the colours

Four colours, and nobody agrees what they mean.

The colour on the bag tells you what the mill calls its setting, not how much fruit was left on.

The names come from what the drying coffee looks like. A little mucilage dries pale and straw-coloured; a lot dries through amber and red to nearly black, because the sugars brown in the sun the way a cut apple browns on a plate.

The main lever is how much mucilage the pulper leaves on.¹⁷ Shade follows from it: pale honeys dry in the open, darker ones are shaded or covered so the drying is drawn out.¹⁹ The speed the water leaves is what actually changes the cup, because it decides how long the sugars and the microbes on the parchment have to work.

Now the awkward part. The percentages are not standard, and neither are the names. One trade guide puts red honey at about 50 percent of the mucilage left on; another puts it at 75.¹⁹²¹ A 2025 survey says outright that "what one producer creates as red might be another producer's yellow or black". ¹⁷ And there is no study behind any of these percentages. I looked, and so did the person writing the sibling post on the dry route, and neither of us found a peer-reviewed measurement of mucilage retained against a colour name, or against anything else. They are trade vocabulary, from roasters and producers rather than research, and the vocabulary is local.

fig 2The same four words, as three trade guides place them. Each dot is one guide's own number, and a line joins the same word from guide to guide. All three meet only at black and disagree through the whole middle, which is where most honey coffee is. Scrub across and the readout says which word each guide puts nearest. Hover, drag or arrow along the axis.trade sources 17, 19, 21 · interactive

So the word tells you the mill's vocabulary and roughly where on the dial it was aiming. Two bags carrying the same word can sit twenty five points apart, and the number that actually changes the cup is almost never printed:

what the label says · what it fixes

{ "process": "red honey",        // the mill's word. Somewhere from 50% to 75% left on.
  "mucilage_retained": null,     // almost never printed, and not standardised
  "drying_days": null,           // the number that actually changes the cup
  "covered": null,               // shade or open sun. Changes it again.
  "turned": null }               // how often, for how long

I do not think the trade is being dishonest. A useful description escaped from a handful of Costa Rican mills and spread faster than any definition could follow it. That is an ordinary thing for a word to do. It is only a problem when the word is printed as though it were a specification.

For the enthusiast: the three guides' numbers, and a contradiction about black honey

The three sets. A 2015 trade guide puts yellow honey at about 25 percent of the mucilage left on, red at about 50 and black at about 100.¹⁹ A roaster's guide from 2016, describing Costa Rican mills specifically, puts the same words a full step apart: yellow at 50, red at 75, black at 100, with a golden honey at 25 in between.²¹ A 2025 survey gives white at about 10, yellow at 25, red at 50 and black at 75 to 100.¹⁷ None of these is wrong; they describe different mills. But they are not defined terms with a body behind them, the way a screen size or a moisture percentage is, and they are not standardised between countries. Every percentage in this post, including the ones in my own figures, is somebody's practice written down.

Three things separate the grades, and only the first is usually on a bag: how much mucilage is left on, how much shade the lot gets, and how fast the water comes out, which is the consequence of the other two.

The disagreement is not only about numbers. The 2015 guide says black honey is covered specifically "to elongate the drying period".¹⁹ The 2016 Costa Rican guide says it is "covered and left to ferment with a relatively quick drying time compared to the other honeys".²¹ Those are opposite claims about the same word, from two credible sources, and neither cites a measurement. A Hawaiian producer says simply that drying times can double between a white and a black honey.²² I have gone with the longer reading in my figures, because it matches every producer account of black honey being the hardest lot to manage, but I am choosing between contradictory trade claims rather than reporting a fact.

03 the difficulty

The hardest of the three to get right.

A seed coated in sugar syrup is the wrong thing to dry evenly outdoors, so honey is the most work and the most risk.

A washed coffee dries as a bare seed in its shell. A natural dries as a whole fruit, which is slow but forgiving, because the skin holds everything together. A honey coffee dries with a coat of syrup on it. Three problems follow, and together they are why honey lots cost more to make than either neighbour.

It sticks. Sticky beans clump, and a clump dries on the outside and stays wet in the middle, so the bed has to be raked several times an hour at the start.¹⁹ That is a person with a rake, not a machine, for most of the first day.

The window is narrow. Wet sugar in a warm place ferments: yeasts and bacteria start feeding on it. A little of that is the fruity, winey note people pay for. Too much and the lot tastes sour, boozy or of vinegar, and is worth less than a plain washed lot would have been. The producer is steering between two failures with a rake.

The weather decides. The method is described as suited to dry countries, because a sticky bean has to dry fast enough not to ferment.²⁰ A run of wet days in the middle of a black honey lot and there is nothing to be done.

And it takes longer. In the sun a washed lot is dry in 8 to 10 days and a natural can take up to 4 weeks.²⁹ Honey sits in between, and nobody publishes days per grade; the one relation I found is that a black honey can take twice as long as a white.²²

fig 3The same cherry, five routes, one clock in days. The ends are sourced: sun-dried washed parchment is off in 8 to 10 days, and a natural takes up to 4 weeks ²⁹. Each honey ends somewhere in the hatch, in order, and black can take about twice as long as white ²². Scrub along the days and the readout says where each lot is. Hover, drag or arrow along the days.FAO ends, trade middle · interactive

So honey asks the most of the producer and leaves the least margin for a mistake. It uses far less water than a washed mill, which is why a lot of people adopted it, but it swaps a water problem for a labour and weather problem. In a good dry season that is a good trade. In a wet one it is not.

For the enthusiast: the raking schedule, the defect risk, and which post owns what

The schedule. One trade guide describes raking or agitating the coffee several times an hour for the first six to ten hours, and then turning it at least once a day for six to eight days after that.¹⁹

The defect risk. The 2015 guide puts it plainly: the longer the coffee is dried, the more likely it is to ferment and develop defects.¹⁹ What is actually happening inside that fermentation, which microbes arrive when and what they make, belongs to the fermentation post.

The weather. Pulped natural is described as suited to countries with low humidity, because coffee covered in mucilage has to dry quickly enough to avoid fermenting.²⁰ The method is a bet on the forecast. The spoilage risks of long outdoor drying, including mould and the toxin that can follow it, are covered in the natural process post, which owns that ground.

04 where it came from

Brazil made it. Costa Rica named it.

The method is Brazilian and older than the word honey; Costa Rica's small mills gave it the colours.

Brazilian mills call it cereja descascada, peeled cherry, and in English it is usually pulped natural.²⁰ It was pioneered there in the 1990s for practical reasons: coffee with the skin off takes less room on the drying patio.²⁰ Costa Rica turned it into a vocabulary after 2008, when water restrictions gave producers a reason to stop washing.¹⁷ Its micro-mills, farms doing their own pulping and drying, gave them the equipment. Once a producer owns the pulper, the setting is theirs to choose, and they need words for the choices. The colours are those words. The names travelled; the definitions did not.

For the enthusiast: why it spread, and who uses which words

Brazil. Pulped natural means the skin comes off and the coffee dries with almost all of the pulp still on.²⁰ Brazil is mostly warm, mostly dry at harvest and mostly flat enough to dry coffee on large patios, and a skin-off lot takes up less room on the patio and in the mechanical driers, which is a straightforward saving.²⁰ It also added sweetness and body to coffee grown at altitudes that, by specialty standards, are low.

Costa Rica. The water restrictions followed a major earthquake in 2008.¹⁷ A micro-mill is a farm doing its own pulping and drying instead of selling cherry to a big cooperative mill; the part micro-mills played is my reading of the story, not something that source says.

Why it spread. It saves water, which matters more every decade. It needs no new equipment beyond a pulper a washed mill already has. It gives a producer a second and a third product from the same trees, which is a real advantage when a buyer wants a range. And it lands on a flavour a lot of drinkers like and could not get before: more sweetness and weight than a washed coffee without the heavy, sometimes muddy fruit of a natural.

Who says what. Brazil mostly still says pulped natural and does not use the colours. Central America uses the colours. Ethiopia, Colombia, Panama and Indonesia have all adopted honey processing and each uses the vocabulary a bit differently. There was never a moment where anybody sat down and wrote the standard, and there is no body whose job it would be.

05 what it tastes of

Between the two, and measurably different from both.

Honey lots are chemically different from washed and natural ones. Nobody has shown they are better.

The trade's short version is more sweetness and body than a washed coffee and more clarity than a natural. In my experience that is broadly right. A good red honey tastes like a washed coffee that somebody turned the volume up on: the same shape, with a rounder middle and a sugary finish that lingers. A black honey drifts towards the natural side, heavier and jammier, sometimes with a boozy edge.

That is a taste note, though, and taste notes are the cheapest thing in coffee. What has been measured is narrower. Five laboratories compared honey lots with washed and natural lots from the same harvest, and found different sugars, acids, fats and aroma compounds.¹²³⁴⁵ Every one found that the routes are different. Not one of them shows that honey is better in general, and the differences in cup score, the tasters' marks, are small. The route reliably changes what is in the bean; whether that is an improvement is a matter of what you like.

For the enthusiast: the five studies, and what reaches the seed

Study	Kind	What it compared	What it found
Aswathi et al. 2023¹	peer reviewed	Robusta in India: honey route 192 h against 72 h for the washed control	Different carbohydrates, organic acids and flavour precursors, two compounds identified in coffee for the first time; sweet, tea-rose and chocolate notes the washed lots lacked
Vasileiou et al. 2026²	peer reviewed	Red Catuaí in Greece: yellow and red honey against double washed and slow fermentation, varying the mucilage itself	Retention steers the bean between burning its own fats and building sugars; the washed lot less acidic, 146.67 against 193.33 mL of titrant per 100 g
Wang et al. 2025³	peer reviewed	China: natural, washed and honey green beans, fats only	510 lipids detected, 37 usable as markers of which route the coffee took
Yulianti et al. 2023⁴	peer reviewed	Indonesia: natural, honey and fully washed, green and roasted, a different instrument	Six marker compounds unique to honey in green beans, five more after roasting
Li et al. 2026⁵	peer reviewed	Yunnan: washed, honey and sun-dried, brewed	Twenty three aroma-active compounds, confirmed by rebuilding the aroma from its parts and leaving compounds out one at a time

Better? The robusta study does call its honey lots a value addition that could fetch a better premium.¹ That is one species, one site and one harvest, and none of the others makes the claim.

What reaches the seed. The mucilage does not get into the seed and become flavour directly. It sits on the outside of the parchment while microbes work on it and the whole thing slowly dries, and what reaches the seed is whatever crosses that barrier. How much crosses, and whether the changes are microbial products moving inward or the seed's own metabolism responding, is the fermentation post's question, and it is genuinely open.¹⁵ This post stops at the bed.

06 the same dial, further

Everything called experimental is this dial, turned harder.

Once a producer owns the fruit, the vessel and the clock, honey process is just the first thing they can do with them.

The honey route taught producers that the fruit is an ingredient, and that they can decide how much of it to use. Everything filed under experimental processing is a further answer to that question. Here is what is done and what is claimed. What the microbes are doing belongs to the fermentation post; I am deliberately not explaining it here, because I would get it slightly wrong and that post gets it right.

sealed tankCoffee shut in a closed vessel, where the fruit's own gas pushes the air out. Usually called anaerobic, without oxygen. Trials at 24, 48 and 72 hours all still cupped as specialty.⁹

carbonic macerationWhole cherries sealed under carbon dioxide, borrowed from the way Beaujolais is made. In one trial on a single lot, the sealed runs beat the open ones and the unfermented control.²⁷

pitched cultureA chosen yeast or bacterium added at the start, the way a brewer does. Also called inoculation. The longest run of research of anything here, some of it on the honey route itself.²⁶

thermal shockHot water then cold, or a hot rest then a cold one, between stages. Widely sold. I could not find a peer-reviewed trial of it, and I am saying so rather than filling the gap.

extended fermentationHolding the coffee longer before drying it. The honey route is itself one: 192 hours against 72 for the washed control in one comparison.¹

co-fermentationSomething from outside the coffee goes in the tank: fruit pulp, juice, cascara, cinnamon, cacao. Called infusion when it is a concentrate or an extract. This is the one where the argument is real.

fig 4The practices on the one axis they share: how long the fruit is held. Every dot is a duration a published trial actually ran; the hollow ones are two Colombian sealed runs. Producer practice goes much further. Two rows are empty because I could not find a trial to draw. Hover, drag or arrow along the hours.published trials · interactive

The shape is the point. The measured region is small: the published work clusters in the first few days, at a handful of sites, mostly in Brazil. Past that edge is real, often skilled practice, with no measured curve behind it. The field's own 2024 review says the new methods "lack sufficient scientific evidence". ¹²

That is not a reason to dismiss the practice. Cooking ran ahead of food chemistry for several thousand years. It is a reason to be careful with the word because. A producer who says "I held this at one temperature for four days and it tastes like this" is reporting. A catalogue that says "the anaerobic environment concentrates the fruit sugars, which is why it tastes of blueberry" has invented a mechanism to sell a bag.

For the enthusiast: the trials behind each practice, with their numbers

Sealed tank. Pulped coffee or whole cherries; the fermentation finishes with very little oxygen present.⁹ Producers claim more intense fruit. The measured result is a different microbial population and a different chemical profile from an open tank, across several regions and both the natural and the pulped natural routes.⁸¹⁶

Carbonic maceration. Carbon dioxide is pumped in, or allowed to build up, to replace the air; in winemaking it is an old and respected technique. One Brazilian group ran it from 24 to 120 hours at 18, 28 and 38 degrees,⁷ and found the best cup score at 96 hours and 38 degrees.⁶ A Colombian trial in 2025 is the cleanest comparison I found, because it ran five treatments on one lot: no fermentation, 24 and 48 hours open to the air, and 24 and 60 hours sealed under carbon dioxide. The sealed runs scored 86.90 and 86.50, against 82.15 for the unfermented control and 84.70 and 85.10 for the open ones.²⁷ Producers claim wine-like and floral notes.

Pitched culture. A 2014 study inoculated four yeast strains into semi-dry, meaning pulped natural, fermentations; DNA fingerprinting showed the added yeast was present throughout, no butyric or propionic acid turned up in any treatment, and the panel found a caramel note the uninoculated control lacked.²⁶ A 2021 study ran four yeasts through natural and pulped natural processing for 27 hours, and the best gave 84.75 and 84.92 points, against controls that also cupped as specialty.¹⁰ Adding a bacterium alongside the yeast is now being tried too.¹¹

Extended fermentation. Beyond a few days the published curve runs out, and what happens next is producer judgement. The published work also covers only a handful of varieties.

07 craft or gimmick

Is this craft, or is it a gimmick.

Most of it is craft. Adding fruit is flavouring. And the real gimmick is the confidence of the numbers.

This is the question the whole category is arguing about, and it deserves the strongest version of both sides before anyone's opinion, including mine.

The case that it is craft

Every method is already an intervention. Honey is a washed coffee with one step left out. The line between processing and interfering was drawn by habit, not principle.
It is the only lever most producers own. Altitude, soil and trees are fixed this harvest; processing is not.
It measurably works. Added cultures take over the vessel, and sealed tanks change the chemistry and the score.²⁶²⁷
Wine settled this argument decades ago. Nobody calls a Beaujolais fraudulent.
The money reaches the farm. The premium goes to the person who did the extra work.

The case that it is a gimmick

Added fruit is added flavour, and the label does not say so. No regulator requires it to be declared.¹⁸
The evidence is thinner than the marketing. The field's own review says so.¹²
It makes origin unreadable. If the loudest thing in the cup is an additive, you have bought a flavour you could have bought anywhere.
The risk sits on the producer. A failed experiment is not a bad quarter, it is a harvest.²³
The rules cannot be enforced. Competitions allow yeast and ban fruit, and no routine test can check.²⁴

For the enthusiast: both cases in full, with their evidence

Craft 1. There is nothing natural about the natural process. A washed coffee is a seed that a person stripped in a machine, fermented in a concrete tank for a chosen number of hours, rinsed in a channel and dried on a bed under a chosen amount of shade. The line has moved before.

Craft 2. Changing variety means tearing out trees and waiting years for new ones to bear. Processing is the one thing a farmer can change this harvest, with equipment they mostly already have, and it can lift a coffee that geography would otherwise cap. For a low-grown farm it can be the difference between a commodity price and a specialty one.

Craft 3. A pitched culture really does take over the vessel, confirmed by tracking its DNA through the fermentation, and really does shift the cup, in pulped natural fermentations as well as in tanks.²⁶¹⁰ Sealing the vessel really does beat leaving it open: five treatments on one Colombian lot, and the two sealed runs scored above both open ones and well above the unfermented control.²⁷ A sealed tank produces a different microbial population and a different set of compounds from an open one.⁸ A systematic review of twenty one studies found 340 distinct volatile compounds across fermented arabica and concluded that induced fermentation, and co-inoculation in particular, widens the range of aromas available.¹³ This is a technique with a literature, not a rumour.

Craft 4. Carbonic maceration is a named, taught, respected winemaking technique, and chosen yeasts are standard in most of the wine world. Coffee is a younger craft holding itself to a purity standard that older ones abandoned once they understood what they were doing.

Craft 5. Experimental lots sell at multiples of commodity prices, and unlike a certification premium the extra goes to the grower. In a crop where the grower usually captures the smallest share of the retail price, that is not a small argument.

Gimmick 1. A coffee fermented with passion fruit pulp tastes of passion fruit because passion fruit was in the tank. When that arrives in a bag labelled coffee, with an origin and an altitude and a variety on it, the buyer is being told a story about a place by a flavour that came from somewhere else. Such coffee is classified as ordinary green coffee.¹⁸

Gimmick 2. Even the carbonic maceration group, after finding their optimum, later wrote that the reason for the improvement remains uncertain, because several of the compounds that correlate with the score are not microbial products at all.⁷ Claims outrun measurements by a long way, and the claims are what is printed.

Gimmick 3. The whole apparatus of specialty coffee, the altitude on the bag, the variety, the region, the producer's name, exists because those things were supposed to be audible in the cup. If an additive is the loudest thing, none of that information is doing any work any more.

Gimmick 4. The buyers who ask for exotic lots carry none of the risk, and the smallholders most likely to gamble on it are the ones least able to absorb a loss.²³

Gimmick 5. The competitions' line is yeast and enzymes yes, fruit and spices no.²⁴ It is a sensible line. A rule that depends entirely on the honesty of the person it constrains is a norm, not a rule.

What the competitions have actually done.

The institutions have been forced to write the line down, and watching them do it is more informative than any catalogue. Cup of Excellence, the auction competition that most defines what a producing country's best coffee is worth, added an Experimental category in Brazil in November 2023: yeasts allowed, whether natural or added, nothing else but water, and fruit and spices banned outright.²⁴ In Costa Rica in 2025, where the three kinds were judged side by side, the experimental winner sold for the least.²⁵

Costa Rica 2025 · top washed: $80.30/lb
Top honey or natural: $91.00/lb
Top experimental: $70.10/lb

One result from one country in one year proves nothing on its own, but it is a useful corrective to the idea that experimental processing is what wins now. In the room where the money is, it did not.

For the enthusiast: the rule texts, the barista championship, and how much of this coffee exists

Cup of Excellence. The Experimental category came more than two decades after the competition began in Brazil, and the executive director's explanation was about fairness rather than principle: comparing apples to apples, and giving every process a fair shot.²⁴ By 2025 the structure had settled into three categories in several countries. Costa Rica ran Traditional Washed; Traditional Semi-Washed, Honey and Natural; and Experimental, defined as processes "differentiated from the traditional" and "produced in a biochemically natural way". The traditional natural won the auction and the experimental lot came third of three. Four of the ten lots in the honey and natural category were honey lots.²⁵

The World Barista Championship went the other way and then partly back. Its rule used to disqualify any coffee with additives, flavourings or aromatic substances added at any point between picking and the cup. It was rewritten for 2024 to forbid additives only after the coffee reaches the green stage, which permits co-fermentation and infusion during processing and bans it afterwards. In the event, no competitor at the 2024 championship used such a coffee. Best of Panama went the opposite way for the same year and excluded infused coffees explicitly, to protect what its organisers called the authentic identity of the sector.²³

The volume. One trade estimate put global production of co-fermented and infused coffee at roughly ten thousand bags, about 0.005 percent of the annual harvest.¹⁸ That is an industry estimate rather than a measurement, it is a few years old, and it is worth holding in mind against the volume of the argument.

Where the evidence is actually thin.

Four things are unknown, and they are the four that matter most. Nobody has published a way to tell a co-fermented cup from an infused one, or from a naturally exuberant one, by taste. Nobody has separated the processing premium from the producer premium. Nobody has shown that any of it holds up in a warehouse. And nobody has measured whether an experimental lot expresses a place better than a conventional one, or covers it up.

fig 5Ten claims, each in the column for the kind of evidence behind it, from not studied to measured again. The mark says the same thing at a glance: a solid slab is measured, an outline is only said, a dotted ghost is unstudied. The column is the kind of source, not my opinion of the claim, and the order inside a column means nothing. Hover, drag or arrow across the columns.sourced · interactive

For the enthusiast: the one laboratory test, and why the rest is open

Taste. The closest work is a 2026 study that built a chemical fingerprint separating fermented, washed and artificially flavoured coffees, and framed it as a question of authenticity rather than of preference. It worked. The fermented samples carried total esters at 74.5 milligrams per kilogram and 2-phenylethanol, which smells of rose and honey, at 27.5; the flavoured ones carried added vanillin at 21.5, had markedly fewer roasting products, and their aroma faded to 55 percent after a fortnight while the fermented markers held.¹⁴ That paper exists because the market made it necessary. It is one laboratory, one study, and a set of instruments no competition has on site: a promising start on a test that does not yet exist.

The premium. The farms winning with experimental lots are, on the whole, the farms that win anyway. No published work runs the counterfactual.

The warehouse. Green coffee is shipped and stored for months, and the one relevant finding, that added flavour faded by nearly half in fourteen days, is about roasted beans, not green ones.¹⁴

The place. That last one is a claim about what a cup is of, and nobody has worked out how to put a number on it.

Where I land.

The dial framing dissolves most of this argument. Sealed tanks, carbonic maceration, pitched cultures, long holds and every honey grade are the same class of thing: they change what the coffee's own sugars do, under conditions the producer chose. Objecting to them on purity grounds means objecting to the washed process too, and nobody does, because the washed process is old. Age is not an argument.

So here is the line I use, and it is about provenance, where the flavour came from, rather than taste. A processing choice is craft if it changes what the coffee's own microbes do with the coffee's own sugars, so that the result is still an expression of that fruit from that farm, however strange it tastes. It is a flavouring, whatever the bag calls it, if the compound you smell came from something added that was not coffee. Both can be delicious. Only one of them is telling you about a place.

That puts black honey, a sixty-hour sealed tank,²⁷ a carbonic maceration at 38 degrees⁶ and a pitched wine yeast on the craft side, and passion fruit pulp on the other. Co-fermentation with added fruit is not fraud and it is not bad, and I have had some I enjoyed. But it should be declared on the bag, plainly, in the same size type as the origin, and judged in its own class. Both are asks about labelling, not bans.

The part I am actually annoyed by is the numbers. A trade guide states that black honey is 100 percent mucilage and another states it is 75, and neither measured anything.¹⁹¹⁷ A catalogue explains that an anaerobic fermentation concentrates sugars, when the study it is loosely gesturing at says the reason is uncertain.⁷ The gimmick, if there is one, is not the fermentation. It is the confidence.

For the enthusiast: the rest of the argument, and six places this post stops being true

The craft side uses the coffee's own fruit, and microbes that came from the farm or from a packet. I am comfortable calling all of that craft, and I think the burden is on the sceptic. In every one of those examples the sugar being eaten came off that farm's own cherries.

The line does not care at all how loud the cup is, which is the usual proxy and a bad one: a well-run natural can taste more extreme than a cautious co-ferment. So co-fermentation with added fruit is a different class of thing, and I do not think the difference is a matter of degree. Flavouring is a category that already exists and already has labelling conventions attached to it everywhere else in food. My position is boring and I think correct.

01The colour grades in my figures are a synthesis of three disagreeing trade guides. There is no authority behind the numbers, including mine.

02The only drying days on this page are the two ends of the dial, from an FAO report. I found no source giving drying days against mucilage retained, and I looked, so the honey grades are drawn relative to each other.

03The processing studies are mostly single-site, single-variety, single-year. They establish that routes differ. They do not establish an ordering.

04Cup-score differences in this literature are mostly small, from one panel on one lot. Treat them as weak evidence.

05The competition results are one country and one year. I use them as a corrective to a story, not as a finding.

06Everything mechanistic is deliberately absent. If you want to know what the microbes are doing, that post is coffee/fermentation and it is the one to read next.

The fermentation is not the gimmick. The confidence is.the sentence this post exists to earn

08 your agent

Use your agent with this page.

Any part of this page can hand your AI assistant a question about exactly that part.

Press ask agent by a heading or a figure caption, or select a passage, and choose Explain it simply, Go deeper or How do we know?. Paste the copied prompt into any agent.

I'm reading "What honey process means". https://hetanshmehta.com/blog/coffee/honey-process Explain the whole post simply, for someone with no science. - With this page's WebMCP tools: call get_outline(), then explain_section({section, level: "child"}) for each section. - If you can open links, read https://hetanshmehta.com/txt/blog/coffee/honey-process.txt. - If neither, say so and I'll paste the post in. Stick to what the page says, and tell me where it stops.

start_guided_readinga lesson, a quiz or the short answer

get_outlinethe map of the post

explain_sectionany section, for a child, a student or an expert

explain_figureany figure, the same three ways

get_evidenceany claim, traced to its source and what kind of source it is

get_glossaryevery technical word in one plain sentence

For the enthusiast: how the tools work

A browser agent with WebMCP asks this page directly; any other agent reads the plain-text version. This subject has more marketing attached to it than measurement, which makes it the worst possible one to read at a single level. A fourteen year old wants to know what the sticky stuff is. A food scientist wants to know which trial, at what temperature, with what control. Somebody's uncle wants to know whether the expensive bag is worth it. The explanations at each level are written by hand, not generated on the spot. The tool that matters most here is get_evidence: every numbered claim resolves to the study behind it, what that study actually measured, and whether it is peer reviewed, trade material or a competition record. Trade claims are marked as trade even when they carry a number. On this subject that last field is the whole game.

This post is a page with tools.

They are registered with WebMCP when the browser has it, which today means behind a flag, and the page does nothing at all when it does not. All six are read-only, all six answer from this page's own data, and none of them makes a network call or needs a key. Their answers load only when an agent first asks.

A plain-text version of the whole post, prose and sources, lives at a stable URL for agents without a browser.

start_guided_readingget_outlineexplain_sectionexplain_figureget_evidenceget_glossaryhoney-process.txt

> get_evidence({ claim: "black-honey-days" }) { "claim": "black-honey-days", "says": "Drying times can double between a white and a black honey.", "kind": "trade", "peer_reviewed": false, "measured": "Nothing. A producer and roaster guide. No measurement method, no sample, no range given.", "source": { "number": 22, ... } }

hover, tap or focus · each one marked peer reviewed, trade, competition record or agency report

Sources · 29 references

Nineteen peer-reviewed papers, seven trade or industry articles, two competition records and one agency report. On this subject the kind of source matters as much as the claim, so every entry is tagged. Where a number comes only from trade practice, the prose says so as well.

peer reviewedAswathi, K.N., Shirke, A., Praveen, A., Chaudhari, S.R. and Murthy, P.S. (2023). Pulped natural/honey robusta coffee fermentation metabolites, physico-chemical and sensory profiles. Food Chemistry, 429, 136897. doi.org/10.1016/j.foodchem.2023.136897
peer reviewedVasileiou, E., Varka, E.M., Nenadis, N., Pavlidou, E. and Paraskevopoulou, A. (2026). Influence of post-harvest treatments on chemical composition, physicochemical, microstructural, and thermal properties of green specialty coffee beans. Food Chemistry, 510, 148665. doi.org/10.1016/j.foodchem.2026.148665
peer reviewedWang, Y., Wang, X., Du, P., Liu, X., He, S., Li, L., Liu, X. and Chen, Z. (2025). Lipidomic profiling provides insights on Arabica coffee flavor diversity in different postharvest processing methods. Current Research in Food Science, 10, 101108. doi.org/10.1016/j.crfs.2025.101108
peer reviewedYulianti, Y. and colleagues (2023). Detection of markers in green beans and roasted beans of Kalosi-Enrekang arabica coffee with different postharvest processing using LC-MS/MS. International Journal of Food Science, 2023, 6696808. doi.org/10.1155/2023/6696808
peer reviewedLi, Z., Gao, Z., Ni, L., Lu, K., Shang, D., Peng, J., Wang, K., Yan, J., Hu, Y., Liu, X., Fan, F. and Tian, H. (2026). Characteristic flavors of different processed Yunnan specialty coffee after hot brewing: from sensory attributes to molecular recognition mechanism. Food Chemistry, 528, 151050. doi.org/10.1016/j.foodchem.2026.151050
peer reviewedBrioschi Junior, D., Guarçoni, R.C., de Cássia Soares da Silva, M., Veloso, T.G.R., Kasuya, M.C.M., Oliveira, E.C.S., da Luz, J.M.R., Moreira, T.R., Debona, D.G. and Pereira, L.L. (2021). Microbial fermentation affects sensorial, chemical, and microbial profile of coffee under carbonic maceration. Food Chemistry, 342, 128296. doi.org/10.1016/j.foodchem.2020.128296
peer reviewedEntringer, T.L., da Luz, J.M.R., Veloso, T.G.R. and colleagues (2024). Genetic diversity of the fungal community that contributes to the sensory quality of coffee beverage after carbonic maceration and fermentation. 3 Biotech, 14(11), 272. doi.org/10.1007/s13205-024-04099-z
peer reviewedPereira, T.S., Batista, N.N., Santos Pimenta, L.P., Martinez, S.J., Ribeiro, L.S., Oliveira Naves, J.A. and Schwan, R.F. (2022). Self-induced anaerobiosis coffee fermentation: impact on microbial communities, chemical composition and sensory quality of coffee. Food Microbiology, 103, 103962. doi.org/10.1016/j.fm.2021.103962
peer reviewedda Silva Vale, A., Balla, G., Rodrigues, L.R.S., de Carvalho Neto, D.P., Soccol, C.R. and de Melo Pereira, G.V. (2023). Understanding the effects of self-induced anaerobic fermentation on coffee beans quality: microbiological, metabolic, and sensory studies. Foods, 12(1), 37. doi.org/10.3390/foods12010037
peer reviewedBressani, A.P.P., Martinez, S.J., Sarmento, A.B.I., Borém, F.M. and Schwan, R.F. (2021). Influence of yeast inoculation on the quality of fermented coffee (Coffea arabica var. Mundo Novo) processed by natural and pulped natural processes. International Journal of Food Microbiology, 343, 109107. doi.org/10.1016/j.ijfoodmicro.2021.109107
peer reviewedCassimiro, D.M.J., Batista, N.N., Fonseca, H.C., Naves, J.A.O., Dias, D.R. and Schwan, R.F. (2022). Coinoculation of lactic acid bacteria and yeasts increases the quality of wet fermented Arabica coffee. International Journal of Food Microbiology, 369, 109627. doi.org/10.1016/j.ijfoodmicro.2022.109627
peer reviewedCortés, V.H., Bahamón Monje, A.F., Bustos Vanegas, J.D. and Guzmán, N.G. (2024). Challenges in coffee fermentation technologies: bibliometric analysis and critical review. Journal of Food Science and Technology, 61(12), 2223-2234. doi.org/10.1007/s13197-024-06054-5
peer reviewedOliveira, E.M., do Rosário, D.K.A., da Silva, B.D., Almeida, A.C.O., Bernardes, P.C. and Conte-Junior, C.A. (2026). Volatilome of fermented and roasted Coffea arabica in Brazil: a systematic review of the influence of microbiota and fermentative processes. Critical Reviews in Food Science and Nutrition, 66(14), 2779-2790. doi.org/10.1080/10408398.2025.2588659
peer reviewedDuan, S., Dong, J., Liu, S., Yu, L., Li, Y., Yin, H., Fang, C., Du, Z., Chen, Y., Zhou, S., Gong, C., Zhang, Q., Ge, Y., Feng, L., Du, X. and Yang, R. (2026). Chemical and sensory profiling of fermented, washed, and artificially flavored coffee beans: insights into flavour quality, authenticity, and food safety implications. Food Chemistry, 525(Pt 1), 150338. doi.org/10.1016/j.foodchem.2026.150338
peer reviewedRevelo-Romo, D.M., Hurtado Gutiérrez, N.H., Hidalgo Troya, A., Amaya-Gómez, C.V., Flórez-Martínez, D.H., Overmann, J., Villegas Torres, M.F. and González Barrios, A.F. (2025). Omics approaches to explore the coffee fermentation microecosystem and its effects on cup quality. Food Research International, 206, 116035. doi.org/10.1016/j.foodres.2025.116035
peer reviewedMartinez, S.J., Batista, N.N., Bressani, A.P.P., Dias, D.R. and Schwan, R.F. (2022). Molecular, chemical, and sensory attributes fingerprinting of self-induced anaerobic fermented coffees from different altitudes and processing methods. Foods, 11(24), 3945. doi.org/10.3390/foods11243945
tradeWensma, T. (8 January 2025). How honey processing is evolving in specialty coffee. Perfect Daily Grind. Industry article, not peer-reviewed; used for the colour-grade percentages, the statement that they are not standardised, and the Costa Rican adoption story. perfectdailygrind.com
tradeKornman, C. (19 December 2022). Additive fermentation: infused coffee is gaining popularity and sparking industry debate. Roast Magazine, issue 115. Industry article, not peer-reviewed; used for the labelling position, the quoted arguments on both sides, and the production estimate, which is an industry estimate rather than a measurement. roastmagazine.com
tradeLynch, R. (7 September 2015). Honey processing: everything you need to know. Perfect Daily Grind. Industry article, not peer-reviewed; used for the raking and turning schedule, the shade practice per grade, and one of the three conflicting sets of mucilage percentages. perfectdailygrind.com
tradeCosta, B. (3 June 2016). Coffee processing: understanding pulped natural coffee. Perfect Daily Grind. Industry article, not peer-reviewed; used for the Brazilian origin and the name cereja descascada, the patio and drier saving, and the humidity constraint. perfectdailygrind.com
tradeOrigin Coffee Roasters (25 April 2016). A guide to Costa Rica's honey process. Roaster's guide, not peer-reviewed; used for the second of the three conflicting sets of mucilage percentages and for the claim that black honey is covered and dries relatively quickly. origincoffee.co.uk
tradeStewart, K. (updated 16 February 2023). What is red, white and black honey processed coffee? Big Island Coffee Roasters. Producer and roaster guide, not peer-reviewed; used only for the claim that drying times can double between a white and a black honey, and for the over-fermentation risk. bigislandcoffeeroasters.com
tradeZanin, R. (16 September 2024). Infused and co-fermented coffee at the World Barista Championship: what should we expect? Perfect Daily Grind. Industry article, not peer-reviewed; used for the WBC rule 3.1.C rewording, the observation that no 2024 competitor used such a coffee, the Best of Panama exclusion, and the producer-risk argument. perfectdailygrind.com
competition recordBrown, N. (8 November 2023). Cup of Excellence coffee competition adds experimental category in Brazil. Daily Coffee News. Trade reporting of a competition rule change; used for the date, the executive director's quoted reasoning, and the rule that only water and yeasts are permitted. dailycoffeenews.com
competition recordAlliance for Coffee Excellence (2025). Costa Rica 2025 Cup of Excellence: categories, results and auction results. Primary competition record; used for the three category definitions, the winning scores and processes, and the auction prices per pound. allianceforcoffeeexcellence.org
peer reviewedEvangelista, S.R., da Cruz Pedrozo Miguel, M.G., de Souza Cordeiro, C., Silva, C.F., Marques Pinheiro, A.C. and Schwan, R.F. (2014). Inoculation of starter cultures in a semi-dry coffee (Coffea arabica) fermentation process. Food Microbiology, 44, 87-95. doi.org/10.1016/j.fm.2014.05.013
peer reviewedMoncayo-Palacios, M.F., Muñoz-Carvajal, V.H., Largo-Avila, E., Suárez-Rodríguez, C.H. and Garzón-García, A.M. (2025). Impact of coffee cherry fermentation methods on the quality attributes of dry-processed coffee. The Scientific World Journal, 2025, 4278424. doi.org/10.1155/tswj/4278424
peer reviewedBernardes, P.C., Coelho, J.M., Martins, P.M.M. and Schwan, R.F. (2024). Microbial ecology and fermentation of Coffea canephora. Frontiers in Food Science and Technology, 4, 1377226. doi.org/10.3389/frfst.2024.1377226
agency reportMutua, J. (2000). Post-harvest handling and processing of coffee in African countries. Food and Agriculture Organization of the United Nations (FAO), AGSI. Technical report, not peer-reviewed; used only for the two ends of the dial, in section 1, Introduction: sun drying of washed parchment "should take from 8 to 10 days", and drying whole cherries "may take up to 4 weeks". fao.org/4/x6939e

Three things I could not verify and so did not state as fact. First, any peer-reviewed measurement of mucilage retained, either against a colour name or against drying days. Every percentage, and every drying time between the two ends of the dial, is trade convention with no published basis, the names are not standardised between countries, and the two trade sources contradict each other on black honey. Second, any published trial of thermal shock as a processing step, which is why that row of the practices figure is empty. Third, any published work on co-fermentation durations or on telling a co-fermented cup from an infused one by taste; the closest is source 14, which is an instrumental fingerprint rather than a sensory test.

merged to main from coffee/honey-process · 94d4e59 · 24 Sep 2026 · the siblings on this branch are the washed route, the dry route and fermentation← the blog