Guava Longjing Milk Tea: A Bar Spec, Brewing Method and Service Plan

Guava Longjing milk tea fails in two predictable places. The first is the tea: Longjing is a delicate pan-fired green tea, and a rolling boil or a forgotten steep turns it into something closer to boiled spinach than chestnut. The second is the emulsion: guava purée is acidic enough to curdle milk proteins, and the failure shows up as flecks, graininess or a visible break — sometimes an hour into service, after the batch has already been sold.

This guide is written for shop owners, bar leads and anyone who has to write the spec card. It covers what Longjing actually contributes to a milk tea, how to buy leaf you can defend on the menu, how to choose a guava format, and a brewing method anchored in published Longjing and green-tea research. It then gives a scalable 16 oz build, an addition order that keeps the milk stable, and a service-day plan.

One caveat before the numbers: the ratios below are a constructed starting spec, not a lab-tested formula. Guava purées differ in Brix, acidity and colour between suppliers and even between batches, and Longjing grades differ in how fast they turn bitter. Treat the spec as a first draft to be tasted, measured and adjusted against the actual purée and leaf on your shelf.

What Longjing Brings to a Milk Tea (and What It Won’t)

Longjing — Dragon Well — is a pan-roasted green tea from Hangzhou in Zhejiang Province. The leaves are pressed flat against a hot wok during firing, which halts oxidation and fixes the characteristic flat, hand-shaped leaf. The infusion is yellow-green, and the flavour is usually described as sweet, fresh, vegetal and chestnut-like (Longjing tea, Wikipedia).

Compared with black or dark tea, that profile is delicate and low in astringency. In a milk tea it reads as the “green” note — aroma, a clean vegetal lift, a soft nutty finish — rather than as the backbone. If you are building a drink where the tea must carry the cup against milk, sugar and fruit, Longjing is the wrong tool.

Be honest about what survives. Milk and guava will mute the chestnut and umami top notes. The tea’s job in this drink is aroma and a clean finish, not dominance. That has a practical consequence for the spec: because the tea is quiet, the milk-to-tea ratio and the tea’s polyphenol composition are doing most of the work on colour and taste. A review of milk tea research summarising Yang et al. (2022) puts it plainly: the colour and taste of milk tea are primarily influenced by the tea-to-milk ratio as well as the polyphenol composition of the tea (Piyasena 2025). That is why the build below fixes both rather than leaving them to the bar.

Know Your Leaf: West Lake, Xihu and the Labelling Trap

There is more than one definition of Longjing, and the gap between them is where margin and menu honesty get tested. A common definition holds that authentic Longjing must come from Zhejiang; stricter definitions limit it to the villages and plantations around West Lake in Hangzhou. The Chinese national standard GB/T 18650-2008 is tighter still: flat green tea produced outside the Xihu and Qiantang areas of Hangzhou, and outside the Yuezhou area of Shaoxing, cannot be marketed under the “Longjing Tea” label (Hou et al., 1H NMR study).

The same paper documents the obvious consequence: unscrupulous traders mislabel green tea from other regions as Longjing to capture the premium that a protected designation of origin commands. And origin verification is not as simple as tasting harder — in that study, a supervised model still struggled to separate samples from four production regions, with Sichuan and Zhejiang samples overlapping heavily.

So origin claims need paperwork, not palate. A short buying routine:

• Ask for the GB/T 18650-2008 designation and production-area documentation on the lot you are buying.
• Treat village names — Shi Feng, Mei Jia Wu — as style indicators rather than guarantees. They describe places on one lake with different characters, not certificates.
• Know that “Bai Longjing” is not a true Longjing. It is a Bai Pian green tea from Anji, made from a white-tea cultivar, and it is commonly confused with Longjing (Wikipedia).
• Check the leaf itself: flat, smooth, even in size, fresh jade to yellowish green, smelling toasty and nutty rather than hay-like or musty.

None of this means you must buy West Lake leaf for a milk tea. It means that if the menu says Longjing, you should be able to show why.

Choose Your Guava Format

Guava comes to the bar in four commercial formats, and each one changes the spec differently.

Single-strength purée or NFC juice. Roughly 7.7–9 °Brix, with 7.7 °Brix the FDA baseline for 100% single-strength guava (Brothers International). This is the format that gives you opacity, natural pectin viscosity and pink colour straight from the fruit. It also brings its own acidity.

Purée concentrate. Typically 20–25 °Brix, roughly 2.6–3.2-fold. It delivers the same opacity and pink colour with less shipping volume, but you are now dosing a concentrated acid load and need to dilute it deliberately rather than by feel.

Clarified juice concentrate. Around 65 °Brix, de-pectinised and filtered. It gives colour and sweetness control with no pulp and no pectin body — useful if you want a clear or lightly hazy drink, less useful if you want the creamy pink body that reads as guava on the menu.

Syrup and jam. These add sugar first and fruit second. They are the easiest to dose and the hardest to balance, because every gram of syrup moves the sweetener line and you end up re-tuning the whole build.

For pink guava purée specifically, a common commercial specification is 8–11 °Brix, pH 3.50–4.20 and titratable acidity of 0.50–0.80% (Dream Foods Caribe). Those numbers matter twice: the Brix sets your sweetness, and the pH sets your curdle risk, which is covered below.

One more variable: guava cultivars differ measurably in soluble solids, acidity and colour. A Korean study of five cultivars found soluble solids from 10.50 to 13.90 °Brix and significant differences in acidity and flesh colour between varieties (Choi & Shin). A spec built on one supplier’s purée will need re-tuning on another’s — not because either is wrong, but because they are different inputs.

Brew the Longjing Base Without Bitterness

Brewing temperature is a crucial lever on Longjing quality — one of several factors, alongside growing environment, harvest season and processing methods. A 2025 study on premium and standard Longjing found that brewing temperature significantly influences the release of both volatile and non-volatile compounds, and therefore the colour, aroma and taste of the infusion (Deng 2025). Two findings from that work are directly useful at the bar:

• Bitterness and astringency rise with temperature and time, and they rise faster in standard-grade leaf than in premium leaf. The study attributed this to the lower tenderness of standard-grade raw material, which leaches flavonols and flavonol glycosides faster as temperature climbs.
• The proportion of key umami-contributing amino acids decreased as brewing temperature increased — which is part of why a hot steep tastes flatter and harsher at once.

On the extraction side, a Turkish green-tea trial tested 75, 85 and 95 °C across steep times from 1 to 45 minutes. EGCG — the dominant catechin — peaked at 85 °C for 3 minutes at 50.69 mg/100 ml, with the highest sensory scores at that condition; at 75 °C extraction was markedly lower, and at 95 °C the peak came earlier and lower, with the infusion turning bitter and dark over long steeps (PMC4573099).

That points to a moderate temperature with a short, controlled steep — not the hottest water you can pour.

Working method for the bar:

1. Heat water to 80–85 °C. If you have no temperature-controlled kettle, boil and let it stand, or pour into a pitcher to cool.
2. Dose leaf by weight, not by spoon — flat leaves take up a lot of volume and spoon doses vary widely.
3. Use a tea-to-water ratio in the 1:30–1:50 range, consistent with published green-tea milk tea preparation (85–90 °C, 7 minutes, 1:30–1:50, tea-to-milk 2:1) (Foods review, Table 6). For a fruit milk tea, start at the shorter, cooler end of both ranges — 80–85 °C and about 3 minutes — and taste.
4. Strain and chill fast. A slow cool keeps the leaf in contact with hot liquor and keeps extracting.

Verification checkpoint. The liquor should be yellow-green, sweet and chestnut-forward, with no drying grip at the back of the tongue. If it grips, drop the temperature or shorten the steep before you blame the leaf.

Build the 16 oz Spec

The build below is a 16 oz (≈473 mL) finished drink. Volumes are given in millilitres for liquids and grams for the purée, with each component also expressed as a share of finished volume so you can scale by multiplication.

The arithmetic. Each share is component ÷ total. For the tea: 180 ÷ 473 = 0.381, or 38.1%. For the purée: 60 ÷ 473 = 0.127, or 12.7%. The total lands at roughly 470 mL rather than exactly 473 because the ice melts into the drink and the numbers are rounded to bar-friendly doses; a few millilitres either way is within build tolerance.

Assumptions to state on the spec card:

• The purée is at its stated Brix. If you switch to a 20–25 °Brix concentrate, you are dosing roughly three times the sugar and acid per gram, so the dose must come down and the water or tea must go up.
• The tea is brewed to the method above and chilled before assembly.
• Ice is counted as dilution. If you build the drink without ice — for example, a shaken version strained over fresh ice — the tea and milk volumes need to rise to keep the same concentration.

Dial-in variants, as percentage shifts rather than new recipes:

• More tea-forward: move 5 percentage points from milk to brewed tea (tea ≈ 43%, milk ≈ 14%). The drink gets lighter and more aromatic; watch that the pink colour does not wash out.
• More fruit-forward: move 3–4 percentage points from tea to purée (purée ≈ 16%, tea ≈ 34%). This raises both sweetness and acid load, so check the curdle test before committing.

Taste the first batch and record what you changed. The spec is only useful once it matches your ingredients.

Stop the Curdle: Acid, Milk and Polyphenols

The curdle risk is real and predictable, and it is worth understanding why rather than just avoiding it by habit.

In native milk, caseins exist in a micellar structure that keeps them water-soluble. In acidic environments below about pH 4.6, casein proteins curdle, forming a curd (Auestad et al., Advances in Nutrition). Pink guava purée at pH 3.50–4.20 sits squarely inside that range. Add the purée to milk and you are pushing the milk proteins toward their instability point.

Tea polyphenols add a second mechanism. They bind milk caseins, which lowers the free polyphenol content of the drink (Piyasena 2025). That binding is part of why milk tea tastes less astringent than the tea alone — but it also means the protein is already occupied, and a sudden acid load has less room to manoeuvre.

Practical controls:

• Add the guava last, into the milk tea. Pouring purée into the cup first and then adding milk concentrates the acid at the interface and raises the chance of a visible break.
• Keep the purée dose within the spec. The 60 g in the 16 oz build is a starting point; if you push it up for a fruitier drink, re-test stability.
• Avoid pairing high-acid purée with under-buffered milk. If your dairy supplier’s product is already near its stability limit, the guava will find it.
• Consider a buffer or stabiliser only if the supplier’s product supports it. Do not improvise with an ingredient whose dosage and function you cannot document.

Verification step. Before committing a batch, hold a sample for the intended service window — the full time the drink will sit on the counter or in a delivery bag — and check for flecks, graininess or a visible break. A drink that looks fine at minute two and splits at minute forty is a refund, not a spec.

Batch, Hold and Serve Through the Day

Turning the spec into an operation means controlling three things: the tea, the guava component and the clock.

Brewing and chilling. Brew in batches sized to your service window rather than the whole day. Chill fast — an ice bath or a blast chiller — and hold the brewed tea in a sealed, dated container. Longjing fades; a batch brewed at open and served at close will taste dull even if it is safe.

Holding the guava component. Follow the supplier’s instructions for the purée or concentrate you buy. Specific hold times and temperatures belong to that documentation and to your local health department, not to a generic recipe.

The regulatory frame. The FDA Food Code is a model, adopted jurisdiction by jurisdiction, so your local rules govern. Where it applies, refrigerated ready-to-eat time/temperature-control food prepared and held for more than 24 hours must be clearly date-marked with a date by which it must be consumed, sold or discarded, held at 5 °C (41 °F) or less for a maximum of 7 days, with the day of preparation counted as Day 1 (FDA Food Code 2022, § 3-501.17). That is the ceiling for date-marked hold, not a target for how long you should keep a batch.

Allergens. Milk is one of the nine major food allergens recognised by US law, alongside eggs, fish, crustacean shellfish, tree nuts, peanuts, wheat, soybeans and sesame (FDA). Your menu and any packaged labelling must declare it. If you offer a plant-based milk version, check that product’s own allergen declaration separately — soy and tree nut milks carry their own.

Common Mistakes to Avoid

Brewing Longjing at a rolling boil. The result is bitterness and a cooked-vegetable taste. Fix: cool the water to 80–85 °C before it touches the leaf.

Steeping past the point of extraction. Long steeps add grip and dullness rather than more flavour. Fix: taste at 3 minutes and strain; do not leave the leaf in the pot through service.

Pouring guava purée into hot liquid, or in the wrong order. Acid hitting warm milk is the fastest route to a break. Fix: chill the tea first, build the milk tea, then add the guava last.

Buying “Longjing” without production-area documentation. Mislabelling is documented and origin is hard to verify by taste. Fix: ask for the GB/T 18650-2008 designation and lot paperwork before the menu claims the name.

Switching purée suppliers without re-tuning. A new purée with different Brix and acidity changes both sweetness and stability. Fix: run the curdle test and a taste panel on every new lot before it goes into service.

Verify the Cup

Run this checklist on the first batch and on every new purée lot. Each item is an observable check, not a number to hit.

• Colour: even pink, no flecks, no pale separation at the top of the cup.
• Aroma: chestnut and green tea still present above the fruit. If you can only smell guava, the tea is being buried.
• Texture: smooth, no graininess, immediately after build and again at the end of the hold window.
• Taste: fruit and tea both legible, with no drying finish at the back of the tongue.
• Stability: no visible break at the end of service. Hold a sample for the full window and look at it before you commit the batch.

If any check fails, change one variable at a time — steep time, steep temperature, purée dose, addition order — and re-run the checklist.

Reader Questions

Is a lower-grade Longjing acceptable in a milk tea? Yes, with a shorter, cooler steep. The 2025 study found that standard-grade leaf leaches caffeine and catechins faster than premium leaf as brewing temperature rises, so it turns bitter sooner. Brew it at the low end of the range and taste early. You will lose some of the chestnut top note, which the milk and guava would have muted anyway.

Can clarified guava concentrate replace purée? Yes, with trade-offs. You gain colour control and lose pulp and pectin body, so the drink will be thinner and less opaque. Because the concentrate is around 65 °Brix, the dose must come down sharply and the tea or water must go up to keep the same concentration. Re-run the curdle test — the acid load per gram is different.

What if a batch curdles mid-service? Discard it. Then check two things before rebuilding: the order of addition (guava last, into the chilled milk tea) and the purée dose against the spec. Re-test a small sample through the full hold window before you rebuild the batch. Do not try to rescue a split drink by stirring — the curd will not re-dissolve.

How long can I hold the brewed tea and the purée? That depends on your supplier’s documentation and your local health department’s rules. The FDA Food Code’s date-marking provision gives a maximum 7-day window at 5 °C (41 °F) or less for refrigerated ready-to-eat time/temperature-control food held over 24 hours, but quality will fade long before that. Brew to your service window.

Three levers decide whether guava Longjing milk tea works on your menu.

First, a cool, short Longjing steep — 80–85 °C, about 3 minutes, strained and chilled fast. That protects the chestnut and green-tea aroma that justifies the tea’s place in the name.

Second, a guava format chosen for its Brix and acidity, not just its colour. Purée, concentrate, clarified juice and syrup each move the sweetness and the curdle risk differently, and a spec built on one will need re-tuning on another.

Third, an addition order that keeps the milk stable: chill the tea, build the milk tea, add the acidic guava last, and hold a sample through the service window before committing the batch.

The next step is a small trial. Build one 16 oz cup to the spec above, run the verification checklist, and write down every adjustment — steep time, purée grams, sweetener volume. Taste it, measure it, and lock the spec before the drink goes on the menu. A drink that is dialled in on paper and verified in the cup is the only version worth selling.