1. You are talking to an AI
Every character in HeyDarlink is an artificial intelligence, not a person. Her messages are written by a language model, and her photos are synthetic images made with generative tools and selected from a library. A few replies are fixed sentences written by our team in advance, for example the safety messages. No person reads your chats or writes replies in real time. The app tells you that you are chatting with an AI at the start of every chat and again every three hours, and the character says so whenever you ask.
2. What the software decides
- How far a conversation may go, within the level your membership, your location and your consent allow.
- Whether a message shows signs of a crisis or of a minor, and the safety response that follows.
- Which photo from the library best fits your request.
- What a character remembers about you, built from what you told her; you can delete it by deleting the message or your account.
- When the app shows you a membership offer, based on your plan, the conversation and your relationship score; you can switch offers off.
3. Limits
AI can be wrong, forget, repeat itself or say something unexpected. Treat everything a character says as fiction, not fact or advice. If you notice a reply that should not have happened, report it through the Complaints and Content Removal policy.