Privacy Policy

HUMANS is operated by Marathon Systema LLC (Wyoming, United States). For privacy questions or requests, contact privacy@humansconnections.com or use our contact page. We have not appointed a data protection officer, because at our size and in the countries we serve we are not required to. Privacy requests reach a person either way, at privacy@humansconnections.com.

• Account: email, first and last name, date of birth. To create and secure your account (contract).
• Phone number: required to create an account, and only one account per number. We text a code through Twilio, who generates and checks it; we never receive or store the code. To keep out automated and duplicate accounts (contract / legitimate interest).
• Conversation with the AI: the messages you send while building your profile. To build your profile and for safety (consent).
• Profile: the structured psychological understanding the AI builds from your conversation, including sensitive inferences (see below). For matching (consent).
• Photos: the pictures you upload. To represent you to a proposed match (consent).
• Preferences: age range, distance, orientation you are looking for, languages, notifications. To run the service (contract).
• Your decisions: whether you accepted or declined. To run and improve matching (legitimate interest).
• Messages with a connection: what you and the person you connected with write to each other. Stored so the conversation is there when you come back, and readable only by the two of you. We also count them, and note when the last one was, to tell whether a connection has gone quiet, which is one of the things that decides when we introduce you to somebody else. Nobody reads them to do that: it is the count and the timestamp, never the words (contract / legitimate interest).
• Safety signals: reports, abuse detection, crisis referrals (references only), audit records. For safety (legitimate interest / legal duty).
• Approximate location: one point, rounded to about a kilometre before it is stored, plus any wider areas you add by name. For distance filtering (consent). If you let the app read your device location we round it the same way and keep only the rounded point, so we never hold precise GPS.
• How you use the app: which screens you open, how long you stay on them, and the actions you take, such as a decision on an introduction, a message sent, or a plan change. Tied to your account id so we can see where the product is confusing and fix it. It never includes what you type, what a button said, or a recording of your screen (legitimate interest).

To find who truly fits you, the conversation may touch on sensitive topics: your faith, your sexual orientation, your sexual life, and emotional patterns the AI infers. We treat these as sensitive and process them only with your separate, explicit opt-in, which you can withdraw anytime in Settings under Privacy choices. You can always decline to answer any question without losing the service, and you can remove anything from your profile at any time.

Every message you send to the AI is screened automatically for signs that someone is in danger: risk of harm to yourself, a threat to another person, someone who may be under 18, or abuse happening now. Most messages are matched against a word list and nothing else happens. When something is flagged, a second automated check grades it, and a record is kept so a person on our team can look. That record holds a reference to the message and a category. It does not hold what you wrote.

There are things we will not keep even if you tell us. The AI is instructed not to record medical or psychiatric details, explicit sexual content, being a victim of violence, criminal history, or anything identifying another person who never agreed to be in our records. If you ask us to remove something from your profile, we remove it and also delete what was derived from it, including earlier versions and any assessment written while it was there.

HUMANS is not a medical, psychological, or emergency service, and nothing here is advice or treatment. If you are in danger, contact your local emergency number. Our crisis protocol explains what the app does when it detects risk.

Your photos are never public. We store them in a private bucket and serve them through short-lived signed URLs; location metadata is stripped. A photo is shown only to a person we propose you to, together with your first name, from the moment the proposal appears, which happens before either of you accepts. Your last name and contact details are never shown, and chat opens only after you both say yes.

Your conversation and profile are processed by AI, and an automated system decides which proposals you see. You have the right to an explanation and to a human review. Full detail is on our AI transparency page.

Your data is stored and processed on servers in the United States by the providers on our subprocessors page. If you live outside the US, this is an international transfer, and we rely on your explicit consent (and standard contractual clauses where applicable).

We keep your data while your account is active. When you delete your account, your profile, conversation transcripts, photos, embeddings, and matching data are deleted, and your account is removed. Messages you sent to a match are anonymized (the body is replaced with a deletion notice) so the other person’s history stays coherent. Consent and safety records may be retained in anonymized form as evidence where the law requires it.

• Access / export: download everything from Settings, or ask us.
• Rectification: your profile updates as you keep talking with the AI; you can also ask us.
• Deletion: in the app (Settings) or without the app at humansconnections.com/delete-account.
• Objection / human review of automated decisions: see AI transparency.
• Withdraw consent: Settings under Privacy choices, anytime.
• To exercise any right without an account, use our contact page. We respond within 5 business days.

All data is transmitted over TLS. Row-level security is enabled on every table, the API authorizes every request in a service layer, photos use signed URLs, and access is recorded in an audit log. We maintain an incident response process. No system is perfectly secure, but we design for defense in depth.

We do not show ads, do not track you across other apps, and do not sell or share your personal information for advertising. We honor Global Privacy Control signals where they apply.

HUMANS is for adults 18 and over. We apply age assurance and act on any signal that a user is under 18 (see our safety page). We do not knowingly collect data from anyone under 18; if we learn we have, we delete it.

We will notify you of material changes in the app before they take effect. This policy is versioned by its “last updated” date.