AI Disclosure

Every persona in Reheart is generated by artificial intelligence. The persona is built to respond in character as your real partner during a session, but it is software — not a human, not a consciousness, and not the actual person. We disclose this clearly:

• before any of your data is sent to an AI provider;
• at the start of every practice session;
• at regular intervals during a session (at least every 30 minutes), so the disclosure persists through longer conversations.

Persona generation, session responses, embeddings, content moderation, and the optional still-portrait avatar are powered by OpenAI models — including OpenAI image generation for the avatar portrait. We name these providers before your data is transmitted, consistent with platform and legal requirements. Our Privacy Policy lists every sub-processor and what each one receives.

A Reheart persona is built from your memories and the data you upload. It reflects your perspective of your partner — not an objective or complete picture of who they are. AI output can be inaccurate, and how a persona responds does not predict how the real person will behave. Treat every session as practice, not prophecy.

Reheart is intentionally not a companion product. Sessions are finite, they end, and the experience is designed to return you to the real conversation. There is no persistent emotional relationship with the AI, and we do not optimize for time spent in the app. This is both a product principle and a safeguard against the dependency risks seen in companion apps.

Our disclosures are designed to meet the leading AI-transparency regimes, including:

• EU AI Act, Article 50 — the obligation to inform people when they are interacting with an AI system and to mark AI-generated content. We disclose AI interaction up front and on an interval basis.
• California SB 243 (effective January 2026) — companion-chatbot rules requiring clear AI disclosure and safeguards. While Reheart’s bounded-practice design sits outside the companion definition, we adopt SB 243’s disclosure standard regardless.
• New York’s AI Companion law — requiring notice that the user is communicating with AI and crisis-safety referrals. We provide both; see our Safety & Support page.

AI output is checked by an automated moderation system before you see it. In emotionally heavy contexts — particularly breakup and goodbye sessions — certain inputs will trigger a supportive, human-reviewed safety message and crisis resources instead of continued role-play. This is a deliberate guardrail, not a malfunction.

For questions about how the AI works or how we disclose it, contact ai@reheart.app.