Last updated: 18 October 2026
MirrorU helps you improve a message you are writing in WhatsApp. This policy explains exactly what MirrorU accesses, what it sends, and what it does not collect. MirrorU is designed to do as little as possible with your data.
MirrorU uses Android's Accessibility service, to read drafts only in WhatsApp and WhatsApp Business. It observes the foreground app to show or hide its button; it does not read other apps' content. When — and only when — you tap the MirrorU button inside a WhatsApp conversation, MirrorU reads the text you are currently typing in the message field, so it can offer a correction.
When you tap the button, the single draft message you are correcting is sent over a secure (HTTPS) connection to MirrorU's correction service, which uses a third-party AI provider (OpenAI) to generate the suggested wording. Only that draft text is sent. The suggestion is returned to your phone and shown to you.
MirrorU's correction service processes your draft to produce a suggestion and does not retain it afterwards. MirrorU does not keep a history of your messages or corrections, and does not use them to build a profile of you. The AI provider processes the text to return a result under its own terms. OpenAI may retain drafts and generated suggestions for up to 30 days for abuse monitoring. This processing is not ephemeral. OpenAI does not use API content to train models by default. See OpenAI’s data controls.
A small amount of standard, non-identifying technical information (such as a request timestamp or error status) may be processed to operate and secure the service. This is not linked to your identity.
MirrorU is not directed to children and does not knowingly collect data from children.
If this policy changes, we will update this page and the “Last updated” date above.
Questions about this policy or your data: support@mirroru.app.