Early Access to Sibyl - FAQ
Thank you for helping us make Sibyl the best possible tool it can be!
You are using a very early, simplistic version of the app. Please read the below FAQ before starting.
Our full privacy policy is here and terms of use can be found here.
Q: What data do you collect?
A: We only collect what’s necessary to make Sibyl work for you:
Your email (to create your account and provide support)
Anything you choose to share about your experience and emotions
Anonymous app usage data (like session length and feature usage)
Q: How do you keep my data safe?
A: Your data is stored on secure EU-based servers and encrypted both in transit and at rest. Only authorized team members can access it. When you send a message, it is securely transmitted to our AI provider (OpenRouter) and then to an LLM to generate a response. See more about which LLMs we utilize below in “Will my data be used to train AI models?”.
Q: Will you share my data?
A: No. We do not sell your data or share it for marketing.
Our long-term goal is to help improve miscarriage and pregnancy loss care worldwide. In the future, we may use fully anonymized and aggregated data to support research or partnerships that lead to better care—but only with your explicit consent.
Q: Can I delete my data?
A: Yes! You can delete your account anytime by clicking “delete account” in the app or by emailing support@sibyl.care. We’ll permanently remove your data within 30 days.
Q: Is this a medical or therapy tool?
A: No. Sibyl provides emotional support, trusted information approved by health professionals, and reflection tools. It does not replace medical advice, diagnosis, therapy, or treatment. Always consult a healthcare professional for medical or mental health concerns.
Q: Will my data be used to train AI models?
A: Sibyl does not use your conversations to train its own AI models.
During our testing phase, Sibyl delivers AI responses via OpenRouter to third-party models (including Gemini Flash by Google). Data handling and model-training practices are governed by those providers’ policies, and in some cases data may be processed or retained for system operation, safety, or improvement.
If Sibyl develops a proprietary model in the future, we will only use data from users who explicitly opt in. We will update this disclosure as our AI infrastructure evolves.
Q: Who can see my data?
A: Here’s an overview of who sees what.
People at Sibyl
Only Melissa (founder) and a small engineering team have access during the beta.
For improvement (anonymized):
We review chat content without your identity attached through a secure admin tool to improve accuracy, features, gaps in knowledge, and safety.
For technical troubleshooting (restricted):
Engineers may access the secure EU-based backend only if needed to fix a bug or crash.
This view includes account details and conversation data, but we do not review chat history unless required to resolve an issue.
Third-Party Services
A few essential partners process limited data to provide their services:
OpenRouter (US): Processes your messages to generate AI responses. May temporarily retain inputs for safety checks but does not train models on your data.
PostHog (EU): Receives analytics only (clicks, usage). No health or chat data.
Google Cloud Platform (EU): Hosts our encrypted [clarify] servers and databases; infrastructure-level access only.
Q: Will I be asked to give feedback?
Occasionally, yes. As Sibyl evolves, we may invite you to share a short, optional check-in to help us understand what feels supportive, what doesn’t, and what’s missing. These are never required, take less than five minutes, and your experience with Sibyl won’t change if you choose not to participate. Your responses are private and used only to help us build a more thoughtful, human experience.
