Forget Hinge or Bumble. This App Promises a Personal AI Matchmaker

It's another day at work: I open the Three Day Rule app, tucked in an inconspicuous folder on my phone next to Hinge, Bumble, and Feeld. From the preview text bubble, I see that I have three new matches and that all three have begun with the same conversation opener: “Hey Molly! I noticed you enjoy live music too; what's the best concert you've seen recently?” Sigh. Here we go again.

Three Day Rule operated as a white-glove matchmaking service for 15 years prior to launching its app in 2025. The app version of Three Day Rule promises a new approach to swipe-centric online dating. Instead of Tinder or similar apps, which start by presenting users a photo, and, if they're interested, provide further prompts to figure out if you have anything in common, Three Day Rule (commonly abbreviated to TDR) starts with several sessions asking you deep questions about yourself and what you're looking for in a relationship, like a real matchmaker would. Once you've answered about 100 questions over the course of a few days, the app starts sending you potential matches.

Traditional matchmakers are often favored by wealthier or older people, and can cost upward of $10,000. TDR CEO Adam Cohen-Aslatei got into the app game by consulting for Bumble, Raya, and the now-defunct S'More, which was sort of a Love Is Blind–type blind-dating app where potential matches started out with blurry photos that came into focus the longer you talked. The idea behind TDR, which uses AI trained by 60 matchmakers, is to make old-school matchmaking more practical, accessible, and aligned with modern sensibilities by incorporating it into a convenient AI-powered app. The app is free, or $25 per month if you want all of the goodies, like AI coaching and guaranteed matches.

So, armed with my AI dating coach “Tai,” I set out to see if an AI matchmaker would find me the person of my dreams from an alleged 250,000 singles ready to mingle.

Instead of a soulmate, I found out that dating is only fun when you're getting to know real people with real quirks. My matches and I talked through the AI prompts rather than getting to know each other, and soon enough, I wondered if the human beings behind the screens were merely facilitating a connection between our AI bots.

Matchmaking, but Make It Modern (With AI, of Course)

You'll download TDR on the app store (right now, it's only available for iOS), and select either the free version, which allows for unlimited chatting with the AI Matchmaker, date coaching, and limited conversations (but no guaranteed matches); or Premium, which has unlimited: “AI-vetted” matches, AI conversation prompts, personalized introductions, matches, and guaranteed matches. (You can also get a 30-day trial of Premium.)

To begin, you'll fill out basic background information, like you would on any dating site: age, location, and profession. You'll also fill out more personal questions like income and racial or religious preferences. Then you'll go through a long series of questions asked by Tai, your AI matchmaker. The goal, Cohen-Aslatei says, is to get “highly curated” matches based on more than hair color and love of tacos: “It's a connection. We're asking you long-form questions to get long-form answers.”

Tai's questions started out broad, asking things like, “What made you decide to try matchmaking?” and “What type of relationship are you looking for?” You can answer via text or voice; the latter makes it feel a little more like you're with a real matchmaker. After each answer, Tai affirms it and asks a follow-up question, beginning with your hobbies and progressing to marital status and educational level. Because this aims to be like a true matchmaker, there were many questions about race, religion, politics, and physical type. Given that I'm a physically disabled, atheist, liberal person who is open to dating all types of people and genders, Tai kept circling back to ensure I was actually that open and not just saying what I thought would be a neutral answer. This is intentional, Cohen-Aslatei says.