Unison is opening its beta in Telegram today. Until now access came by invitation, in small groups, with manual checks along the way; from today anyone with Telegram and a serious intention can join.
Open beta is not a phrase from a press release. It means the product works and does what it promises, but does not pretend to be finished: the matching core is built, the interface will keep changing, and part of what we have planned is still ahead. Beta users will see those changes first — and, more importantly, will be the first to shape them.
What Unison is
Unison is a dating service that matches people not by profile but by imprinting: the subconscious image of a partner formed in early experience, which then works as an unconscious recognition filter. We have already written in detail about what imprinting is and what the research says; briefly — a person notices “their” face in a fraction of a second, long before being able to explain why this particular person seems attractive.
What it looks like in practice. You rate a few dozen photographs: simply mark who triggers a response, with no justifications and no “right” answers. From those reactions, and from the people you yourself name as significant, your preference profile is assembled. The algorithm then looks for those whose visual image matches that profile — and shows a pair only when the match is mutual.
Unison has no swipes. No public profiles to scroll through, collect or show to friends. No endless feed of faces where the value of the next person drops with every movement of your thumb. There is one strong signal — a subconscious response — and an attempt to find it on the other side.
Why we are opening the beta
A dating service cannot be tested in a vacuum. Matching quality depends on how many real people have gone through the ratings, how they react to the pairs they are shown, and what happens after first contact. With invited groups we saw the essential thing: the mechanics work, and people genuinely recognise “their” faces in the output. But a small sample answers only whether this works at all, not whether it works well.
So we are opening access to everyone. We need more than feedback on design; we need real experience — what happens when tens of thousands of ratings arrive from people of different ages, cities and tastes. That is the volume at which it becomes visible where the algorithm errs and where a match turns out to be accurate.
What already works
Matching by imprinting. Photo ratings, a preference profile, a search for mutual overlap — the whole chain runs inside Telegram, with no app installs and no email sign-up.
Privacy by default. Your profile is never published or indexed: only the algorithm sees it, plus the people you have matched with. We do not display your photographs to other users as a feed and we do not turn dating into a shop window.
Human support. A team stands behind the bot, reads every message and answers on the merits. In a beta that matters more than usual: some questions do not have a ready answer yet, and we would rather say so than reply with a template.
What is not there yet
An honest list of limitations is worth more than a polished announcement, so here it is plainly. We are not shipping iOS or Android apps yet — the entire service lives in Telegram, and that is a deliberate choice for the beta: it lets us change the product daily without waiting for store reviews. There is no web account either.
The main limitation is the size of the base. The algorithm looks for mutual overlap rather than simply “showing similar people”; the fewer people near you, the rarer those pairs are. Early in the beta some cities will wait longer than Moscow or St Petersburg. This is not a bug that code can fix: the base grows with the number of participants, and every new user makes matching more accurate for everyone else.
The interface will change too. Flows in the bot, wording, the order of steps — none of it is final. If something feels awkward, we will most likely agree: we have no reason to defend decisions made in a hurry.
What the first participants get
Early access to features arriving in the next iterations, without waiting for a public launch. A say in what Unison becomes: we collect beta feedback into a single list and prioritise it rather than filing it under “someday”. Special terms at launch, which we will announce to beta users separately, before the public release, and only inside the service.
And finally the important part: a chance to find your person earlier than others. Not because the beta offers magical access to a “better” base — it is smaller than it will be. Because dating starts not with endless choice but with one accurate match.
What we ask of you
Write about what does not work — the bot has a support button, and a human reads every message. Tell us not only about bugs but about impressions: a confusing step, a strange pair in the output, a sense that the algorithm missed — all of it is data the next version is built on.
And about safety. Unison is a service for adults; registration is open from 18. We do not ask and will never ask for card details, SMS codes or passwords, and we never write first from unrelated accounts. First meetings are better held in public places — that is a rule about common sense, not distrust of the other person.
How to join
Open the bot @UnisonDatingBot in Telegram and tap “Start”. Then comes a short series of photo ratings, a few questions about who you are looking for, and your preference profile is ready. It usually takes five to ten minutes, and you can pause at any point: progress is saved.
Once the profile is built, the bot starts looking for matches. If there is no mutual response yet, it will not send “similar” people just to keep the feed busy — better to wait for an accurate match than to receive ten approximate ones. A notification arrives in the chat when a pair is found.
What comes next
The coming months go into three things: match quality, geography, and the conversation after a match. The first is the most important and least visible work: improving profile calibration, learning to filter out random reactions, and verifying that a shown pair really does feel like “their own” to both sides. The second is expanding cities and languages so that waiting does not depend on where you live. The third is what happens after a match: right now it opens a dialogue, and we want that dialogue to start more easily than the standard “hi” into the void.
An open beta is not a test of users’ endurance or a way to save on a launch. It is a format in which a product grows up alongside the people it is built for: we show the work before it becomes flawless, and we take on the obligation to explain what is happening and why. If the idea that dating should not start with a swipe and an endless scroll of profiles feels right to you, you are welcome. We read everything you write.