How we protect women

Every line on this page describes something that is running in the app right now. None of it is a plan or a promise. What we haven’t built is at the bottom, by name, because a safety page that only lists wins is a marketing page.

Before anyone can reach you

• Nobody can message you until they have passed selfie verification. Browsing and liking stay open so the app has people in it, but sending a message is gated. That is the real bot and fake-profile choke point, and it is on by default.
• Verification is a live check against the profile photos. The front camera watches while they look straight at it and do three things picked at random (turn left, turn right, close their eyes, smile), and takes a photo at each. Apple or Google confirms it came from the real app on a real phone. Our own servers check every step again, flag a photo of a photo or a screen, and compare the face with the profile photos. A clean pass verifies on the spot. Anything unclear goes to a person on our team. The computer never turns anyone down.
• Every photo has to contain exactly one face. Pets, objects and memes are rejected, and so are group shots, because we won’t guess which face is yours.
• Once you’re verified, every new profile photo has to match your selfie. Our own servers compare the face in it with your verification selfie, with no limit on how many photos. A photo whose face doesn’t match yours, whether it’s someone else or a face made by AI, is turned away. Before you’re verified, the selfie check compares you with every profile photo, and you don’t show up in anyone’s Discover until it passes. A photo with no clear face, like a full-body shot from a distance, has nothing to compare and goes through. Makeup is fine.
• Caption text banded across a photo is rejected. Meme profiles don’t get in.
• Every photo in a set has to be a different photo, by perceptual fingerprint, so nobody fills six slots with one picture.
• A banned account’s photos cannot be re-uploaded. We keep fingerprints of them, and a near-duplicate counts. Resizing or re-saving a photo does not get around it.
• A banned phone number cannot re-register.
• Contact info in a profile is detected automatically: an @handle, a phone number, or a username after an app’s name or an “add me.” Naming an app is fine. It sees through the dodges: look-alike letters, zero-width characters, spaced-out spelling and digits for letters. A clear attempt bans on the spot. Pointing people to an account (“my insta,” “add me,” “hmu”) gets one warning and the text removed, and a second one bans. “Snap a photo” never gets anybody banned. In chat, a message with a phone number, an email or another app’s name asks “are you sure?” before it sends.
• Sign-up and sign-in from a VPN, proxy or datacenter address are blocked, because that is where scam-farm signups come from. It is not a ban. Turn the VPN off and carry on.
• New and unverified accounts get a tighter likes-per-hour cap than verified ones.
• 18+ only, and it is attested at signup. Any attempt to involve or contact a minor is reported to authorities.

What you control

• Block anyone, from anywhere, instantly and both ways. They can’t see you and you can’t see them. It takes no match and no reason, and they are never told.
• Block your entire contact list in one tap. Everyone in your phone disappears from you and you from them, your ex and your boss included. Your contacts go to our server once, where each number is hashed with a secret salt and the number itself is thrown away. We keep only the hashes, and they can’t be turned back into numbers.
• Who can’t see you. Hide yourself from people by distance from your home, by age, or by a value on their profile. They are never told. They simply don’t see you. This one is part of premium.
• Incognito. Only the people you have liked can see you.
• Pause your profile without deleting anything.
• Your home is a neighborhood, never an address. The GPS fix only decides which census tract you are in, an area the Census Bureau draws around 1,200 to 8,000 people, and is then thrown away; what we store is the middle of that tract. Nobody has a house-level point for you, including us, so working out where someone lives from distances, the trick researchers used on six big dating apps in 2024, ends in a neighborhood of thousands. Your card shows your city, and two people in the same tract see no distance at all.
• Export everything, or delete everything. Export is in Settings and hands you a file. Delete account is on your Profile tab and is permanent.
• Private answers stay private. Shadow answers are never shown on your profile. They only decide whether you match somebody else’s private filter, and nobody ever learns what anyone searched.

Once you are talking

• Abusive first contact never reaches you. Until you have said something back, a like comment or a message carrying a threat, sexual content, a slur or a degrading insult is held before it gets to you. You never see it or get a notification about it, and nobody asks whether it bothers you. The sender is told it was held and a person on our team reviews it. If the reviewer decides it was fine, it is delivered late. Otherwise it stays held for good. Swearing on its own goes through.
• Five conversations at a time, and nobody can buy more. He cannot keep fifteen women on the hook, and he cannot pay to be at the front of your five. A slot only frees when a conversation is actually ended. Going quiet doesn’t do it.
• Somebody at their cap is not shown to you at all, so you are never matching into a queue.
• The app warns before contact info goes out. A phone number, an email or another app in an outgoing message gets a heads-up first, because that is how most scams start. So does a first message the filter above would hold, so the sender can fix it before it goes.
• Photos in chat arrive blurred. You tap to see each one, unless you turn on auto-show in Settings yourself. It is off until you do.
• Video messages cannot be screenshotted. The capture is blocked on iPhone and Android. If somebody tries on an iPhone, a line appears in the thread saying they did.
• Money talk gets flagged, and an ask gets reported. A message that asks for money, gift cards, crypto or a payment app goes to our safety team as a report the moment it’s sent, and a warning with a Report button shows under it on her side. He isn’t told.
• Plan a date sends a friend where and when, his profile and photos, and a code word, before you go. The app asks “Home safe?” when the date should be over, with help and reporting on the same screen. Share for safety on a profile does the same.
• Report and block are on every profile and in every chat, free, no match required.
• Unmatching does not end your ability to report him. Past matches keeps anyone you matched with for 30 days after it ended, either direction. The button that makes it stop can’t double as the button that protects him. It does not say who ended it.

What gets reported, and what doesn’t

Anyone can report anyone. It needs no match and no explanation, it costs nothing, and the person you report is never told who did it. We do not make women justify a report before they are allowed to file one.

The server decides how serious it is, from your answers. The app only collects them, so nothing the reported person does can talk a report down. There are three tiers.

What is not reportable

Worth being just as clear about, because a queue full of nothing is a queue that misses the real thing. Being unmatched, ghosted or turned down is not reportable, and there is no button for it. Neither is a profile you find unattractive, an opinion you disagree with, or someone at their conversation cap who hasn’t replied. Unmatch, or use the filters.

Weapons in a photo are allowed. A range day or a hunt in somebody’s photos is normal here and banning it would make nobody safer. Explicit and threatening content is handled on its own merits.

What happens after you report

• It lands in a queue a person works through, and a person writes the answer.
• Three reports against one account, from anyone, flags that account for high-priority review and bumps every open report against it.
• Every report closes with an action and a written reason: dismissed, a warning, content removed, account hidden, or account banned. Bans take effect immediately.
• A catfish report can carry photos you took. They go straight to storage, never through the app’s own servers in the open. Only our safety team can open them, through a link that expires, and they are deleted when the report closes. The person you reported never sees them and they never appear anywhere in the app.
• A ban record keeps hashes, never raw data: a hash of the phone number and fingerprints of the photos. A leak of it would reveal no numbers and no images, and it is what stops a banned man starting over.
• Every ban can be appealed in writing, and a person reads it. If we got it wrong you are back the next time you open the app.

What we never do

• We never sell your data, and there is no advertising or cross-app tracking SDK in the app.
• We never tell someone you reported or blocked them.
• We never charge for safety. Blocking, reporting, contact blocking, the incoming filter, crisis resources and seeing everyone who likes you are free, and they stay free.
• We never sell a way around the rules. You cannot buy a sixth conversation and you cannot buy your way to the front of somebody’s five.

What we have not built

This section exists so the rest of the page can be trusted.

• The incoming filter reads words and only screens first contact. It does not look at photos. Once you have replied, the conversation is yours, and report and block are there for anything that turns.
• We do not run background checks, and we cannot guarantee anybody’s identity or conduct. Verification checks that the face in the photos matches the person holding the phone. It does not prove they are safe.
• The automatic selfie check has limits. It can’t reliably catch a live deepfake fed in through a fake camera, an emulator or a tampered phone, or a 3D mask, and no outside lab has certified it. That’s why it can only pass people, and why reports still matter.
• Screenshot blocking covers video messages only, not photos and not profiles.

Pending features

Planned or switched off. Nothing else on this page counts on them. Each one moves up to the lists above once it’s on and we’ve checked it works.

• Detecting AI-generated photos on their own, a face filter on a photo (lenses, a warped face, stickers), and a face that looks like a minor’s so it can be blurred. Built, and switched off on 3 October 2026: the outside service that does them limits how many photos it checks each month. The selfie match above runs on our own servers and has no limit.
• Detecting explicit photos in chat. Every photo you receive is already blurred until you tap it. This would keep an explicit one blurred and labelled even if you’ve turned auto-show on, and flag whoever sent it to our safety team.
• Detecting subtle beauty filters and retouching. Not running. We haven’t found a free tool that does it reliably.
• ID verification, a tier above selfie verification.
• Flagging scam and fake profiles automatically. Today it starts with your reports.
• Keeping a reported explicit image as evidence, so it can go to the police if you want to file. This one waits on our lawyer’s review.

If any of that changes, this page changes with it. If you think something here is wrong, tell us and we will fix it the same day: dev@pappasproductions.com.

How to report or block someone, step by step · Bans and appeals