A Sprout Works concept by CoreyOctober 2026
One palate that tells you what to order, what to make, and what to buy next.
Skim the headlines in two minutes · Flip a card where you want more
Part 1 of 4
Who it's for, why the timing works, and why no one owns cocktails yet.
Part 1 · The opportunityFour people
The same taste profile does a different job for everyone who touches a cocktail.
Going out
Scan a menu and get every drink ranked for your palate in about three seconds, zero-proof included.
Fall menu, ranked for you
5 drinks read in about 3 seconds
Home bartender
See what your bottles make tonight, the near-misses, and the next bottle ranked by taste. Learn why you like what you like.
Your shelf
Bars and bar teams
Opt-in palates, menu-gap insights, menu drops and seasonal passports.
Creators and organizers
Creator collections with revenue share. Trails and cocktail weeks run on one contract. Brands come later, as labeled sponsored content only.
Stirred and bitter
A collection by a cocktail creator, 12 recipes
Cocktail week trail
3 of 6 bars. Finish for a non-alcohol perk.
Part 1 · The opportunityWhy now
People drink less and choose each drink more carefully, with zero-proof growing alongside alcohol. iOS 27 reads images on the device.
Why now, in three signals
Part 1 · The opportunityThe gap
Recipe apps know your shelf, AI scanners read menus, and drinks networks log bottles. No one ties them to one palate with users behind it.
| Capability | Recipe and shelf appsMixel, Cocktail Party, PDT, Cocktail Flow | AI menu scannersStirred Art, Drinkist, DunDunDunalmost no ratings | Drinks networksUntappd, Vivino, Beli | proposed |
|---|---|---|---|---|
| Ranks a bar menu by your taste | ||||
| Shows what your shelf makes tonight | ||||
| Picks the next bottle by taste | ||||
| Logs cocktails you order out | ||||
| A palate that travels between bar, home and store | ||||
| Guest-taste insight for bars |
Does it Partly, or only for wine, beer or venues Not found in the research
Recipe and shelf apps: Mixel, Cocktail Party, PDT, Cocktail FlowAI menu scanners: Stirred Art, Drinkist, DunDunDun (almost no ratings)Drinks networks: Untappd, Vivino, Beli
Groups are marked by their best member. Full marks show capability, not traction.
Why no cocktail log has stuck
No label
A bar's Negroni has no barcode, so there was nothing to scan. A menu scan fixes that.
A chore
Logging while out with friends is friction. Here it falls out of ordering, in one tap.
No reason
Some people "don't see a reason to track drinks."41 Here every rating sharpens your next pick.
Part 2 of 4
How one palate works at the bar, at home and at the store, and how it stays honest.
Part 2 · The productHow it works
One palate answers three questions: what to order at the bar, what to make at home, and what to buy at the store.
1
At a partner bar, the code at your seat opens its menu, already loaded. You choose whether the bartender sees your palate for that visit. Anywhere else, point at the menu and on-device vision reads it, house cocktails included.
2
Every drink gets a fit % and a confidence level, zero-proof included, with unprofiled house drinks marked as estimates. Order from your picks through the bar's own point-of-sale.
Tap a drink on the phone to see why.
3
"You ordered the Oaxacan Negroni. How was it?" Three colors, then an optional head-to-head. Each rating nudges your palate.
Try it on the phone.
4
One tap if the bar's menu is already on file. A photo of the menu if it isn't. Or type the ingredients, with autocomplete. Every save adds that bar's version to the city's catalog.
Tap the three ways on the phone.
5
One photo stocks it. Nudge a bottle's level if the photo can't tell, then see what it makes tonight and what's one bottle away.
Tap bottles on the phone to add or remove them. Illustrative sample of 11.
6
Ranked by how well it fits your palate, then the recipes it unlocks and the price. It follows your shelf.
7
Chat with your shelf, with a menu, or about why you like what you like. Short answers, grounded in your palate and your bottles.
A direction, not a commitment. Tap a question on the phone.
8
Your shape, your house pours, and a taste match with friends.
9
Every drink you rate shows who made it. Bartenders keep a pro profile with their palate, signatures and favorite bartenders. Follow them, or shout out a drink they made; they and the bar see it.
A proposed direction. Tap Follow or Shout out on the phone.
Fall menu, ranked for you
Menu on file at this bar, seat 4
YouThe drinkConfidence
Bitter and stirred like the Negronis you rate highest, with the smoke you rated up last month.
Smoky enough for you, but shaken and sourer than you usually pick. Its profile is an AI estimate from the recipe; no bartender has checked it yet.
Spirit-forward like you like, but sweeter and less bitter than your usual.
The best zero-proof pick here: herbal and bitter-edged, without the spirit.
Sweet and fruity, the opposite corner of your shape.
Your pick: the Oaxacan Negroni
After your round
You ordered the Oaxacan Negroni. How was it?
Better than last week's Paper Plane?
Saved. Your palate moved toward smoky.
Save a drink you had
Saved
Oaxacan Negroni
at an example bar, tonight
From the bar's menu on file. One tap.
Add an ingredient
mez
Each save also adds this bar's version to the city's drink catalog. Illustrative.
Your mixologist
I have rye, Campari and a lemon. What can I make that's not too sweet?
A rye sour with half an ounce of Campari: tart and bitter, light on sugar. Pick up dry vermouth and you can make an Old Pal.
Which of these is closest to a Paper Plane?
Low Tide. Aperol and citrus like a Paper Plane, with mezcal smoke in place of bourbon. 71% for you.
Why do I like mezcal drinks?
You rate bitter, stirred drinks highest, and mezcal adds smoke without adding sweetness. Your favorite mezcal drinks all sit on that side of your shape.
A direction, not a commitment. Illustrative.
Your palate
Updated after tonight's Oaxacan Negroni
NowLast month
Your shelf
Tap a bottle to add or remove it
Illustrative: 11 bottles, 13 recipes; assumes citrus, sugar, honey and ice.
Next bottle
Ranked by taste fit, then recipes unlocked and price
Fit is the best match among drinks that use the bottle. Illustrative.
YouA friend
Bitter and stirred, with some smoke
House pours
Taste match with a friend85%
Maya R.Pro
Bar lead at an example bar. Verified by the bar.
Palate
Bright and herbal, a little smoke
Signatures
Bartenders Maya rates
From guests
YouBest Oaxacan Negroni I’ve had. Thanks, Maya.
JessThe zero-proof Garden Party is why I came back.
Illustrative screens. Drink profiles and match numbers are demo values.
Part 2 · The productPick five
Choose one of two classics, five times, and watch your shape form. Then compare it with a few drinks, including a house cocktail nobody has rated.
Pick 1 of 5
Which would you rather have?
Go with your gut. Pick the one you'd order.
Five picks done
Here is your palate.
Try it against a drink
Low Tide
mezcal, Aperol, grapefruit, salt
Likely a fit
Your palateThe selected drink
Demo profiles are illustrative; real profiles come from the bartender panel.
How the palate learns
Recipes
Vivino's match needs crowd ratings of the exact bottle.1 A cocktail is a recipe, so it can be placed before anyone rates it.
Bartenders
Bartenders score what each ingredient adds. AI drafts a new drink, labeled an estimate until a person verifies it.
Your ratings
The real model learns from your ratings over time. This demo uses rough draft profiles. illustrative
Part 2 · The productBuilt with bartenders
Every profile traces back to people who make drinks for a living. AI places new drinks, clearly marked as estimates, until a bar team or a trusted taster verifies them.
Bartenders
Bartenders score how each ingredient pulls on the six axes.
A tasting panel
Credited bartenders taste and score the classics. Their names stay on the profile.
AI, clearly marked
AI places a new drink from its recipe, with a confidence level. It stays "estimated" until verified.
Estimated, low confidence
The bar team or trusted tasters
The bar's own team or trusted tasters adjust the profile and earn the drink a badge.
Verified by the bar team
You
Every rating you give tunes your own matches over time.
In the open
Predicted matches are checked against the ratings people actually give, and the results are shared.
One drink, before and after a bartender tastes it
Low Tide an example bar's house drink
mezcal, Aperol, grapefruit, salt
AI estimate · low confidence
Profile confidenceLowPlaced from the recipe alone.
Shown as an estimate until the bar team or trusted tasters verify it.
Illustrative. Demo profiles are illustrative; real profiles come from the bartender panel.
Trusted tasters
Verified mixologists and bar managers keep the profiles honest.
Fit and confidence, kept apart
Fit is the chance you'd rate a drink four stars or higher, learned from real ratings. Confidence is how sure we are of that number, shown as a range. A strong fit we're unsure of says so, instead of hiding inside one score.
92% fit
High confidence
A verified drink and plenty of your ratings. Likely 88 to 95%.
84% fit
Medium confidence
A verified drink, but only a few of your ratings so far. Likely 70 to 93%.
Likely a fit
Low confidence
A new house drink, estimated from its recipe. It could land anywhere from 50 to 95%, so no single number.
Illustrative. Confidence comes from three things: whether a bartender has verified the drink, how many of your ratings back the match, and how many people have rated the drink.
When tasters disagree
Median
A profile uses the median score, so one outlier can't swing it.
Re-taste
A wide spread between tasters triggers a re-taste, not an average.
Bar's call
On its own house drink, the bar team has the final say, and the profile says so.
Proposed. Who sits on the tasting panel, and how trusted tasters get verified, is still to be decided.
Part 2 · The productWhich feels right?
Three directions that differ in who talks to you, what you see first and what happens at the bar. The palate, data and guardrails stay the same. Nothing is decided.
Your palate
Bitter and stirred, with some smoke.
It moved toward smoky after last week’s Oaxacan Negroni. Scan a seat or a menu when you’re ready.
Oaxacan Negroni
How was it?
Your shape shifted a little toward smoky.
Only in refined
Your spec book.
1 oz mezcal
1 oz bitter aperitivo
1 oz sweet vermouth
Stirred, on the rocks, orange peel
Quiet by default: no notifications unless you turn them on.
Concept, illustrative
Nobody talks on screen. The palate is the only character, with a bartender’s short notes.
Point me at a table card, or ask me anything. I know what you like.
Ask your drop…
Smoky suits you. Called it.
Next time’s dare
A zero-proof you’ve never triedEarns a stamp. Variety counts, never volume.Your drops, side by side
85% taste match
Sam leans sour and herbal. Try a Last Word together.
Dares done
Concept, illustrative
Your drop, a character shaped and colored by your palate, talks you through the night. Dry wit, never a baby face.
Behind the bar at an example bar
“I’d pour you the Low Tide. Smoky like you like, but brighter.”
Maya
A signed pick, matched to your palate.
Low Tide, made by Maya
How was it?
House pours
Favorite bartenders
JessThe zero-proof Garden Party is why I came back.
The barThank you! Maya has a winter one coming.
Concept, illustrative
Real bartenders talk, through pro profiles and signed picks. The drop is only a small cameo.
Hybrid is the concept lab’s recommendation; Corey hasn’t picked. The drop is an original placeholder sketch drawn to read adult: it never appears at the age gate or paywall and never cheers another round. Bartenders are fictional.
Part 3 of 4
How every scan builds a catalog, and how guests, bars and creators connect through it.
Part 3 · The networkThe catalog
Every scan or saved drink records a specific bar's drink, ingredients, price and date. Vivino built its catalog from label scans. Cocktails never had a label, so scans and saves build the catalog.
Part 3 · The networkPro profiles
Verified bartenders and bar managers keep a pro profile: their palate, their signature drinks and the bartenders they admire. Guests follow them and shout out what they loved. Proposed for after v1.
Brand-paid posts are labeled #ad, and rewards are never drinks. Bars can reply to posts and feature them with the guest's OK, but can't hide or buy reviews.76
Part 3 · The networkFor bars
Guests opt in per venue, revocable anytime. Scanning the code at a seat shows the bartender that guest's palate for the visit. That one consent powers a recommendation at the bar, menu-gap insights, a seasonal passport and menu drops. Guests' public shout-outs about your drinks and team land here too.
Venue dashboard, an example bar
Illustrative dataWho's drinking here
The palate mix of guests who opted in this season, shown only above 25 guests.
One opted-in guestTheir average
Your guests lean that way. Nothing on the fall menu reaches it.
Spec idea, drafted from your back bar
1½ oz mezcal
¾ oz bittersweet amaro
¾ oz sweet vermouth
2 dashes mole bitters
Stirred, rocks, orange peel
Predicted match with opted-in guests: 71%, estimated until your team tastes it
How each drink matched and rated
Match with opted-in guests, and the share who rated it "loved it."
What guests are saying
Public shout-outs that tag your bar, your drinks or your team, in one place. Reply as the bar, or feature a post on your menu and socials once the guest says yes.
JessThe zero-proof Garden Party is why I came back.
DevLow Tide is smokier than the menu says. In a good way.
SamLong wait at the bar on Friday, but worth it.
Fall passport
A stamp per drink tried. Finishing earns a seat at the bartender's class, funded by the bar. Zero-proof counts, stamps are capped per visit, and the window is the whole season.
One guest's passport, 6 of 8
Publish a menu drop
Reach guests within
Publishing a menu is free.
380nearby users whose palate matches 3+ drinks on this menu
now
Three drinks on this winter menu fit you.
Why a bar would want this
63%
of guests who already know their order can be swayed by a bartender's recommendation.11
$50/mo
buys POS loyalty today, so retention is covered.12 A POS can't find new guests whose taste fits the menu, so we plug into it: picks and orders flow into systems like Toast (proposed).
25+
opted-in guests before any insight shows, in aggregate only. Individual palates are never sold.
Part 3 · The networkCreators and organizers
Creators sell collections to the palates they fit. Organizers run trails and cocktail weeks on one contract. Brands come later, labeled.
Recipes ranked for each reader's palate, with age checks and disclosure handled by the app.
Organizers run the whole week on one contract, with taste-matched discovery for every participant.
Labeled, pre-approved sponsored content only, once there's an audience to show it to.
Part 4 of 4
How it makes money, stays compliant, and launches.
Part 4 · The businessThe money
Scanning, logging and the palate stay free; depth costs about $29.99 a year. Year two adds campaigns for organizers and a venue tier sold on palate insights.
First
Scan, log and palate free
~$29.99 a year
~$69.99 lifetime
Year 2
~$100 per venue per campaign, plus an organizer license
$49–99 a month venue tier, sold on palate insights
Last
Labeled sponsored content only
Comparable: Untappd for Business, about 20,000 venues48,60 at $899–1,199 a year.61 All prices here are proposals.
The benchmarks
Part 4 · The businessCompliance
Compliance is designed in from the first screen, so bars, organizers and brands don't have to work around it.
The rules behind it
Federal
A brand may not give a bar anything of value, even through a third party.59 Brand money buys brand-level content only.
State
Some states ban free drinks, drinks as prizes or buy-one-get-one deals.68,69 Rewards are status and non-alcohol perks.
21+
Apple's age-range check blocks minors,73 and the app adds its own 21+ gate.
Research, not legal advice. A beverage-alcohol lawyer reviews the rewards design before any B2B contract.
Part 4 · The businessLaunch
Each phase has a pass mark before the next begins. The first city is chosen with launch partners, then the app grows city by city.
Build the cocktail graph, then test scanning on 50 real menus.
Gate90%+ read correctly, under 3 secondsour threshold
About 100 enthusiasts, recruited from cocktail communities.
Gate40%+ log five drinks in 14 daysour threshold
A set of bars, one cocktail-week organizer and a couple of creators in one city, chosen with launch partners.
GatePartners signed; menus live; a first trail or weekproposed
Expand once a metro reaches density, then open the venue tier there.
GateAbout 2,000 monthly active loggers per metroour estimate
Go-to-market channels
Creators' audiencesLaunch collections reach their viewers.
Weeks and trailsOrganizers bring many venues at once.
Table QR codesBars invite guests to opt in.
App Store searchRatings compound over time.
Why each gate
90%
of drinks read right, in under 3 seconds. No public benchmark exists for dim, stylized menus, so the scan is proven first. our threshold
5 drinks
logged in 14 days by 40% of testers. Five mirrors Vivino's five-rating start.1 our threshold
~2,000
monthly loggers per metro before selling to bars, since bars won't join an empty app. our estimate
End notesSources
Every market figure links to its source. Labels mark everything else:
The palate shape has six axes: spirit-forward (stirred) at the top, sour (shaken) at the bottom, bitter opposite sweet, smoky opposite herbal.
Part 3 · The networkSocial
Social starts with identity, not a feed.
Logging networks grow through what people show about themselves: a palate card, house pours, an annual recap, taste matches with friends, and city rankings of single drinks.
Palate card
YouA friend
Bitter and stirred, with a little smoke
85% taste match with a friendIllustrative
What the analogs show
Letterboxd
30M+
members by July 2026, up from about 1.8M in 2020
Logged898.5M
Rated672.5M
Reviewed143.6M
Listed12.9M
2025 activity. Most people log and rate; few publish.49
Beli
80%+
of users joined through referrals
New users invited four friends to get in, with an eight-person team.37 As of October 2025 it was "not monetizing in any significant way."50
Untappd Recappd
~12M
year-end recaps in 2025
An annual recap people share, with a brand sponsor.51