How does Catina differ from uploading health data to ChatGPT?
A regular ChatGPT conversation uses OpenAI's hosted consumer service and follows your current training, retention, feedback, and safety controls. Health in ChatGPT has additional protections for connected health information. Catina instead stores supported records in an encrypted local Health Vault and uses browser-to-attested-enclave encryption for current sensitive AI request and response bodies. Operational metadata still remains visible where needed to run the service. Read the current comparison.
How does Catina compare with Health in ChatGPT?
Catina is a privacy-oriented health intelligence platform for importing lab reports from any lab or country, combining them with supported genetics and lifestyle context, tracking biomarkers, and generating personalized insights from an encrypted local Health Vault. Health in ChatGPT is a provider-hosted option for eligible US adults who want Apple Health, supported medical records, Function Health, and a mainstream assistant. OpenAI says connected health information is excluded from foundation-model training and ad targeting. Read the current comparison.
Do you store my health data?
Supported records are stored in an encrypted local Health Vault. Standard unlock is backend-assisted. Extra Local Protection separately adds a passphrase-gated local encryption layer. Sensitive AI features can send encrypted request bodies to an attested enclave for processing; this is separate from vault storage. We recommend keeping an encrypted export and the original source files.
Who can access my data?
Access depends on the feature. Your device holds the supported vault record. Standard unlock involves Catina's account backend, while Extra Local Protection adds a passphrase-derived layer. On supported sensitive AI paths, plaintext request bodies are intended to be available inside the verified enclave, while the relay and network providers can process operational headers and metadata.
What is a TEE and why does it matter?
A Trusted Execution Environment is a hardware-isolated processing environment that can produce attestation evidence about the code and configuration it is running. Catina's browser client verifies the selected enclave before establishing the supported encrypted body path. Device integrity, operational metadata, hardware trust, and supporting services remain part of the complete trust model.
Can my data be used to train AI models?
Personal Health Vault records and protected sensitive AI bodies are excluded from model training. Decryption happens exclusively inside the verified enclave. Account metadata, support messages, public community content, and other product surfaces have separate purposes and policies.
Is my data sold or used for ads?
Personal Health Vault records are excluded from sale and targeted advertising. Catina is funded by paid plans. Service providers receive limited account, payment, infrastructure, support, analytics, or encrypted-processing data as needed for their functions.
Can I sign up anonymously?
Yes. A made-up username is enough to create an account. Providing your real name, location, and other identifying details is optional. Any email address you provide is used for account access and support.