Three openers in your voice. Fifteen seconds.
Paste the profile. Get a quick read on it, plus three things you'd actually say — ranked by who's most likely to reply.
What we actually return
Not a personality essay. A quick read — what's working in their profile, what's a question mark, and what hook is most likely to land. You can scan it in the time it takes the elevator to reach your floor.
green flags
- mentions her sister twice — close family
- specific Sunday plans (farmers market, no brunch) — direct
- answered the prompt earnestly, no bit
yellow flags
- every photo is a group photo — hard to tell which one she is
- no prompts about what she's looking for
best hook
the line about her grandfather's pasta sauce — she'll talk about that for 20 minutes if you let her
Calibrate your voice once. Three texts.
Paste three messages you've actually sent — the kind you'd send to a friend at 11pm. We learn how you talk: do you use periods or not, do you double-text, are you the dry-jokes person or the earnest person, do you ever use "lol" unironically. After that, every opener comes out sounding like you wrote it on a good day.
The next day, we ask: did they reply?
One tap. Yes or no. Over a few weeks, we learn which openers actually work for you — not for the average user, for you. The model tunes to your real reply rate, not a hypothetical one.
Generic AI opener vs. yours
The difference between something that gets left on read and something that gets a reply at 11:47pm.
What happens to your screenshots
Three things, plainly. One: every screenshot you upload is deleted within 24 hours, automatically, no setting required. Two: nothing you upload is ever used to train models — not ours, not anyone's. Three: voice calibration lives encrypted on your account; you can wipe it in one tap from settings.
Pricing
Free tier covers most weeks if you're matching at a normal pace. Paid is for when you're actually dating.