Why we built Cohort

What felt broken

Dating often rewards the shallow snapshot: a fraction of a second to judge a face, then another, then another. It trains inconsistency and leaves a lot of people quietly drained - across every gender. Attention skews hard: a small slice of profiles absorbs most of the interest, while many others barely get a fair read.

Too many apps make money when you stay in motion but never quite land. Their economics can quietly depend on keeping people from actually connecting. That is not a foundation we wanted to stand on.

North star

We want a product that pushes back on the halo effect - the shortcut where one striking trait washes out everything else - and instead creates room for true connection: honest, deep, and meaningful enough to respect your time.

Why cohort, spark, thread, recap, contract - not swipe, then chat

Swipe feeds optimize for volume and impulse. Chat opens before either person knows what they are actually signing up for. That sequence maximizes messages, not mutual comprehension.

Cohort starts by placing you in a small weekly group aligned on intent, so your environment matches how seriously you are dating right now. A spark is scarce on purpose: it forces intent and cuts performative breadth. When a spark is mutual, you enter a guided thread - timed steps where you respond in sync and reveal answers in pairs, so neither person is performing into a void.

Voice shows up inside that thread when it belongs to the conversation - not as a novelty clip on a card. That keeps tone and language tied to context, which matters when you are deciding if someone feels safe and intelligible to you.

The recap makes the implicit explicit: here is what we actually said and chose together. The contract is where you both state whether you want to continue - and on what pace - before photos and open chat unlock. It is reciprocity made legible: clarity before intimacy scales, not after.

What we believe about dating products

• Depth over raw volume. Adding options faster than people can metabolize them does not increase good outcomes - it increases noise and regret.
• Reciprocity is signal processing. Connection is not a broadcast of preference; it is verified, bidirectional bandwidth. Products should make mutual intent easier to read, not harder.
• Ambiguity is structural debt. What stays fuzzy tends to compound into misunderstandings. Small costs to clarity upfront save larger costs later.
• Slowness can be a feature. Sampling life at a sustainable rate preserves nuance that disappears under infinite scroll.
• Safety is design, not decoration. Friction that respects pacing is different from surveillance theater - it is how nervous systems stay in the conversation.

What we reject

We do not want to profit by preventing people from connecting. If the business model ever quietly rewards loneliness loops, we have missed the point.

How we decided to be different

The basis is not one viral feature - it is a bundle of choices: separate intent lanes so mixed motives do not collide in one deck; scarce sparks so attention behaves less like spray-and-pray; guided steps with paired reveals so first impressions can revise instead of freezing at the halo; recap and contract so ambiguity is treated as a bug, not a retention trick. Marketing pages describe the shape of that journey; we avoid publishing internal prompt text so the experience stays fair for members.

Tradeoffs we accept

Cohort is slower than infinite swipe, offers fewer simultaneous options by design, and asks for patience with invites and cohort timing. If your priority is maximal browsing speed or unstructured chat with as many people as possible, mainstream apps will fit better - and that is a reasonable preference.

Where we are rolling out

We are starting in Berlin, expanding across Germany, then the United States, with a path toward serving people globally. This site is in English so early members everywhere can read how we work before they join.

Invites

Access is invite-only while we keep cohorts small and the experience coherent. Every request is read by a real person before placement - not to crown a narrow idea of who is "interesting," but to confirm people understand Cohort's pace and intent model so weekly groups stay workable for everyone inside them.

Request an invite when you are ready to try it.