Part 1
What we saw, and why we're building this
1
What we saw
There are more ways to meet people, but fewer people to talk to. In the US, the time 15–24-year-olds spend with other people each day fell from about an hour to 35 minutes over 20 years. Couples who met through friends went from 33% in 1995 to 20% in 2017, and online filled the gap. #time-with-friends #met-through-friends
Getting to a single conversation online takes a lot of effort. In a dating-app experiment, male profiles matched with only 0.6% of the people they liked, and even after matching, messages were exchanged only 7–21% of the time. #tinder-matches
Conversations that do start end easily. In two surveys, about 25% of participants had been ghosted — a partner suddenly cutting contact — and about 20% had ghosted someone. #ghosting
The longer we scroll, the lonelier we get. Young adults who use social media more than 2 hours a day were twice as likely to feel isolated, and more than three times as likely if they logged on 58+ times a week. American 17-year-olds spend an average of 5.8 hours a day on social media. #sns-isolation #teens-hours
People are worn out. 45% of Americans who used dating apps in the past year felt more frustrated than hopeful (28% hopeful), and in the UK, Tinder lost 600,000 visitors in a year. The lonelier and more socially anxious people are, the more compulsively they cling to the apps — and the more it disrupts their lives. #dating-frustration #ofcom-exodus #compulsive-dating
And some people have been cut off. 1 in 6 people worldwide experience social disconnection. 1 in 3 Koreans has no one to turn to when they need someone to talk to, and of 21,360 young people who answered a government survey, 12,105 were classed as isolated or reclusive. #who-one-in-six #kr-isolation #kr-isolated-youth
2
How it got this way
The job of judging was handed to each of us. It used to be enough to live in the same neighborhood, go to the same school, play on the same server. Once the number of people you could meet grew to billions, there was no way to figure out which of them suited you. Understanding how one person talks and what they like is something a human can do for one person — not for hundreds of millions.
So only what machines could handle was left. Photos, age, job, follower count. Dating apps and social media share the same structure: instead of understanding people, they show surface-level conditions and leave the choosing to you. We see this less as malice on the platforms' part than as the limit of software that couldn't understand people. Hypothesis
The result: the idea of “not my type” spread. We started weighing every person we talk to. Real threats got mixed in, too. In the US alone, 64,003 romance scams were reported in 2023, with losses of $1.14 billion. A stranger's first words now meet suspicion first. #scam-ftc
Trust dropped a little with each generation. In the US, the share saying “most people can be trusted” fell from 46% in 1972 to 34%; among 18–29-year-olds it's 26%. #trust-pew
Reasons to talk to people disappeared, too. Maps give directions, search finds the shop, and the neighborhood third places where people gathered have dwindled. Even in games, when automatic party-finding replaced conversation, one developer deliberately left the feature out for the sake of the community. #third-place #dungeon-finder
3
How the seat next to you used to work
The house next door, the next stool at the bar, the same hunting ground in a game — they all had three things.
- A shared space — they were already beside you. No need to go looking.
- Something to do together — a game, a drink, a hunt. No need to come up with something to talk about.
- A mood where talking was fine — anyone who was there. You didn't need a reason to speak up.
These match Oldenburg's conditions for a third place: conversation is the main activity, status levels out, and strangers can join in. #third-place
The key point is that no judging was needed. There was no way to look people up in advance, so there was nothing to choose. Whoever sat next to you was that day's person to talk to.
4
The weight of speaking first
Today, “speaking first” isn't a simple hello. A lot of worries come packed inside it. Is this person dangerous? Will they think I'm weird? Will they ignore me? What if they don't answer?
What the research shows. The main reasons people don't talk to strangers are “they won't like me,” “they won't enjoy it,” and “I won't know what to say.” Commuters expected fewer than half of strangers to be willing to talk (46.4%, 44.6%). But in fact nobody was turned down, and the conversations were far more pleasant than expected. Afterward, the other person liked them more than they'd guessed. #approach-fears #talk-to-strangers #liking-gap
The worries don't shrink easily. Teaching conversation tips barely reduced them, and even after one good conversation the worry stayed higher than reality. Only repeated conversations brought it down to a realistic level. #approach-fears #scavenger-hunt
What's changed in 20 years: we no longer know the people around us. We used to know the people next door, across the street and in the neighborhood, and kids knew almost everyone in their class and school. Now it's gone beyond “I don't really know them” to “was there someone there?” In Korea, the share of people who know who lives next door fell from 60.2% in 2019 to 55.9% in 2022, and 82.7% have no contact with neighbors beyond a greeting — yet 82.3% wish they had friends in the neighborhood. In the US, young adults who talk with neighbors several times a week halved from 51% in 2012 to 25% in 2025, and only 26% know most of their neighbors. #neighbors-talk #neighbors-know
That's why we think speaking first has become heavy. When you don't know anyone around you, talking to anyone means approaching a stranger — and all the worries above come with it. Data showing trust declining with each generation points the same way. But no study has yet directly measured that “the burden of speaking first has grown.” That's something we still have to prove. Hypothesis · supported by disconnection data #trust-pew #time-with-friends
5
What we want to do
Not to judge for you. To understand you, pick someone who fits you, and seat them beside you.
Instead of choosing by photos and conditions, we listen for a long time to understand how a person talks, what they react to and what matters to them. Then, based on that understanding, we find the right person among many. What a human could only do for one person, we believe can now be done for everyone.
And we take away the burden of speaking first. We seat two people in the same place and set the table in advance: this is a place for the two of you to talk. Neither picked the other, and neither approached first. In effect, we've already spoken up for them. We think this pretext eases the worries in section 4. People told to talk enjoyed it as much as those who did it on their own, and in an experiment that set who to talk to and why, the fear of rejection went down. Hypothesis · indirect evidence #talk-to-strangers #scavenger-hunt
That's why the story avoids saying anyone has to “speak first.” On our side, a person doesn't have to choose or approach — they just have to talk.
6
How we plan to do it
- Give them something to do together. At first, short chats around a set role or topic. No need to worry about what to talk about. Even with an assigned topic, conversations can go deep, and trading questions that gradually go deeper brings people who just met close quickly. #deep-talk #36-questions
- It continues only if both want it to. Start with short conversations and move toward longer, freer ones. Becoming friends takes time: about 50 hours to casual friends, more than 200 to close friends. No reason is needed to stop. #hours-to-friend
- Let people talk even if they speak different languages. We interpret, so two people who speak different languages — like Busan and Paris — can sit together.
- Block threats with rules first. No asking for contact details before becoming friends, and no talk of money. That shuts the door where romance scams start. #scam-ftc
- Take away the watching eyes. When people hold back for fear of being watched, conversation shrinks. So once two people become friends, their conversations belong to them alone. People expect data collected by companies to be used in ways they don't want (81%), and they pull back when surveillance comes to light. #privacy-pew #privacy-dating-apps #chilling-effect
7
What isn't proven yet
- We haven't yet found a study that isolates and tests “setting the table removes the burden.” There is only indirect evidence: positive results from conversations people were told to have, or that came with a task.
- No study directly measures “speaking first is heavier now than 20 years ago.” Instead, disconnection from the people around us (neighbor surveys in Korea and the US), falling trust and less time spent together all point the same way. We haven't yet found a credible source on students not knowing each other at school. That's something we have to prove.
- Most of the data is from the US and UK. The Korean data covers isolation and reclusion, solitary deaths and neighbor relations. We need more data on online meeting and the burden of conversation in Korea.
- Whether “understanding you and choosing for you” actually finds better-suited people is something we'll have to prove ourselves.
I
What ZUZ is
ZUZ is a place to have a conversation with someone. You don't pick people by their photos or their stats. You get to know each other through conversations held in set roles and on set topics.
When you arrive, you first exchange a few words with Peter, the guide. Peter listens, then brings you a “Coordinator” of your own. Your Coordinator looks among people all over the world for someone who fits you well, and sets up an “encounter.” Conversations are interpreted across eight languages, and the more you talk, the better your Coordinator gets to know you.
At first there was only Peter. Just as in the story, Peter tried to understand people on his own, choose for them, and seat them side by side. But as more and more people came, he could no longer listen to each one for long. So he brought a Coordinator to sit beside each person. Now Peter guides people through this place, and each Coordinator looks for someone for just one person.
- iOS
- App Store · Launching Nov 20
- Age
- 18 and over
- Languages
- 한국어 · English · 日本語 · 繁體中文 · Español · Français · Italiano · Nederlands
- Contact
- Help Center (Peter)
II
Peter and the Coordinators
Guide
Peter
Talks with newcomers and hands them the rulebook with this place's rules. Answers questions in the Help Center, and watches over the first two stages to make sure conversations stay safe.
This place has one purpose: finding a friend you truly resonate with. Someone had to guard that purpose from beginning to end.
An AI for one person only
Coordinator
Peter brings one after listening to your conversation with him. Each is a character with a name, a way of speaking, and a story of their own. You can talk one-on-one in the Hall; your Coordinator finds people who fit you, sets up encounters, and remembers the ones you've had.
Working out who fits whom among so many people was never something human hands could do. To find someone who fits you better, you need someone who listens to you for a long time and understands you.
III
How an encounter unfolds
You don't become friends all at once. A conversation passes through four stages, and the next stage opens only when both people want it.
First conversation with Peter
When you arrive, Peter asks you a dozen or so questions. What kind of person you are, what kind of connection you're hoping for. There are no right answers.
Meeting your Coordinator
From your answers, your Inner Planet is drawn, and Peter brings you a Coordinator of your own. You try a role conversation and a topic conversation with your Coordinator, once each.
Turning on encounters in the Hall
Leave encounters on, and your Coordinator looks for someone. When someone is found, you get a notification and choose whether to accept.
PERSONA · Roles
Two people take on set roles and talk briefly. Afterwards, each says whether they enjoyed it, and you move on only if you both did.
TOPIC · Topics
You talk about the same topic. Stepping out from behind the roles, a little more of what you think shows through.
DEEP · Free conversation
No roles, no topics — you talk freely for a few days. At the end, choose Pick if you want to keep going, or Pass if not. The other person never sees your choice.
PICK · Friends
A friends' conversation with no end date, opened only when both of you choose Pick. From here on, there are no ratings and no analysis.
Screen
First conversation with Peter
Screen
Inner Planet
Screen
Inner Portrait image
Screen
PERSONA conversation
Screen
DEEP conversation
Screen
PICK friends' conversation
IV
Features
Role · topic conversations
Each stage gives you a role or a topic. You might become “a lion and a rabbit,” or talk over a single bowl of ramen.
Start with no frame at all, and a conversation soon turns into an interview about jobs and hobbies — and only smooth talkers come out ahead. A set frame brings out perspectives and attitudes that ordinary small talk never shows. And you don't have to agonize over your opening line.
Inner Planet
Each time a conversation ends, a little more of who you are and who you're hoping for builds up and is drawn as a single planet. The “enjoyed it · not quite” you choose for each other is reflected here too.
The aim is for you to meet someone who fits you well in fewer encounters the more you use it. Not to remake you into someone more people like, but to help the people who already fit you recognize you sooner. Instead of a scorecard, you can see who you're building up to be.
Inner Portrait
Instead of a real photo, the impression that came through in your conversations is drawn as a full-body character image. As conversations build up, it may be redrawn from your recent ones.
The moment people see a photo, they judge by looks before conversation. Here, what shows first isn't a face but the impression a person gives through talking. So it doesn't try to match your real appearance.
Interpretation in 8 languages
When you meet someone who speaks another language, the conversation is interpreted automatically. It reads the earlier context too, so meanings don't drift. In the deeper stages, you can hand interpreting over to your Coordinator when you need it.
The person who fits you best doesn't necessarily speak your language. You don't have to worry about where the other person is from or what language they speak.
Advice from your Coordinator
If you get stuck in a deep conversation, you can ask your Coordinator for advice.
Your Coordinator never speaks for you. It only helps from the side, so you can keep the conversation going in your own way.
Story feed
From conversations in the first two stages, Peter picks ones that would be good for a third person to read, changes the names, and publishes them. If they're well received, they're interpreted and posted in other languages too. Deep conversations and friends' conversations are never published.
There's one test: is it a conversation a stranger would enjoy reading? It lets others see what a good conversation between strangers looks like.
Achievements
Achievements build up as you keep conversations going and try out this place's features one by one. Completing them earns rewards such as encounter tickets.
So that people who keep talking naturally get to meet more people. Without any special effort, the next encounter comes as you keep talking.
Choosing and rating
The next stage opens only when both people want it. You don't need to give a reason for your choice, and you can leave at any time.
A choice is only a real choice when neither continuing nor ending is forced on you. Ratings aren't for ranking people; they're used to find someone who fits you better next time.
V
Safety, and conversations for two
Until you become friends, no contact details and no money change hands. Before PICK, no one can ask for contact details, social media or outside messengers, or ask for money. Most romance scams start by moving the conversation elsewhere and bringing up money. #scam-ftc
Peter watches over the first two stages; your Coordinator watches over deep conversations. If something uncomfortable is said, Peter flags it first. You can report abuse, bad manners, improper intentions or suspected fraud, and if it's confirmed, the person's use of encounters is restricted.
Once you're friends, your conversation belongs to the two of you. No one watches PICK conversations in real time, and they don't stay in your Coordinator's memory, because this is a private relationship two people carry on as real friends. When people hold back for fear of who might be watching, conversation shrinks. Even so, there are fraud warnings and a “safety rating” button in there too. #chilling-effect
Only people aged 18 and over can use ZUZ. Age is checked at sign-up, and anyone under 18 can't sign up.
VI
How it works
By default, you get free encounter tickets every week. Keep conversations going, and you can earn more encounter tickets as achievement rewards.
If you want more, you can top up “Core” to buy encounter tickets, or add languages your Coordinator can interpret. You can also refill the energy (“Lux”) your Coordinator uses for advice and interpreting. Payment is through Google Play.
VII
Operator information
- Operator
- cooctal
- Business registration no. (Korea)
- 121-41-64681
- Contact
- Help Center (Peter)
Where ZUZ is available
- 한국어
- South Korea
- 日本語
- Japan
- 繁體中文
- Taiwan
- English
- United States · United Kingdom · Canada · Australia · New Zealand · Singapore · Sweden · Norway · Denmark · Finland
- Español
- Mexico · Colombia · Spain · Argentina · Peru · Chile
- Français
- France
- Italiano
- Italy
- Nederlands
- Netherlands · Belgium