Read the privacy policy of any AI app on your iPhone, the actual policy and not the summary page. You will find that conversations may be stored, reviewed for safety, or used to improve models, depending on the provider and on the settings you choose. Those settings and policies change, so check the current ones before you rely on any of them.
Satcove's Privacy Shield addresses one specific part of that problem: what identifying details reach the AI providers. This article explains what it does, what it does not do, and when it is worth turning on.
What Privacy Shield does
Privacy Shield is an opt-in setting, available on every plan including Free. When it is enabled on a supported text flow, Satcove tries to detect common identifiers in your message, such as names, email addresses, phone numbers, IBANs and ID numbers. It replaces them with neutral placeholders before the text is sent to the AI models, then maps the placeholders back to your real values in the answer you read.
The effect: the models answer your question without receiving the identifying details you typed. If you ask about a contract and your name is in the text, the providers see a placeholder instead.
What Privacy Shield does not do
Being precise about the limits matters more than the feature itself.
- It does not mean your chats are never stored. Privacy Shield does not turn off chat history and does not delete saved conversations. Your history is kept according to your plan and settings.
- Detection is automated and imperfect. It can miss unusual formats or identifying context, for example a rare job title and a small town in the same sentence.
- It covers supported text flows only. Images, audio, attachments and some tools may need separate care. Remove sensitive details you do not need to share.
- It is not anonymity and not compliance. It reduces exposure. Whether your use is lawful still depends on your own notices, retention rules and provider terms.
- It is off by default on the web. You turn it on in Settings, Privacy Shield, for that browser. Check the setting on your iPhone before a sensitive question.
What Satcove says about training
Satcove does not use your conversations to train a Satcove foundation model. The AI providers process the text you send under their own API terms, which apply on top. For the full wording, read the Satcove privacy policy and the Privacy Shield page.
When it is worth turning on
Turn it on when your question contains details that identify you or someone else: a medical situation with names, a dispute with a named employer, account numbers, a contract with party names. For general questions with no personal details, it changes little.
A simple habit also helps: do not type what you do not need. Replace the name with "my employer" yourself, and keep only the facts the question requires.
Reliability is a separate question
Privacy Shield does not make the answers more accurate. The models Satcove queries still have knowledge cutoffs and can be wrong. For health, legal or financial questions, treat the consensus as a starting point and check an official source or a professional.
Get the app on the App Store, then check Settings, Privacy Shield, before your next sensitive question.