Moonpool ("the app," "we," "us") is a personal wellness app for mood check-ins, written reflections, gratitude notes, breathing exercises, movement logging, custom habit tracking, and optional AI-assisted reflections. This policy explains what we collect, why, where it goes, and the choices you have.
The essentials:
- Your data is yours. Synced data is isolated to your account with per-user access rules, protected with TLS in transit, and encrypted at rest.
- Your content is not used to train AI. We use OpenAI and Anthropic commercial APIs without opting in to model training. Provider retention and its exceptions are described below.
- Nothing goes to AI until you say so. AI features are optional and require your consent before first use. After that, sending a message or requesting a reflection in a saved chat can also trigger automatic updates to her memory when remembering across chats is enabled.
- Delete means delete. Deleting an entry erases its content from our backend, not just your screen; deleting your account removes your account and synced content.
- No ads or cross-app tracking. No advertising SDKs, no tracking across other companies' apps or websites, and no advertising identifier.
For analytics and diagnostics, we use TelemetryDeck and Sentry, described below. These services use pseudonymous identifiers: hashed or randomly generated identifiers that can link events from an app installation or device without using your name or email. We do not use advertising SDKs, track you across other companies' apps or websites, or use the Advertising Identifier (IDFA).
What we collect and why
Your wellness content. Mood check-ins (including any emotion words you tag), written reflections, gratitude notes, breathing sessions, movement entries, and any custom habits you create. Your entries are saved on your device, and if you sign in they sync to your private account on our backend so your devices stay in step. If you don't sign in, your entries are not synced. Content can still be sent for AI features you have consented to or included in exports you create.
Account information (only if you sign in). Signing in is optional. If you sign in with Sign in with Apple or an emailed one-time code, we receive your email address and an account identifier so we can create your account and sync your data across your devices. Sign in with Apple lets you choose to hide your email, in which case we receive Apple's private relay address.
A device-scoped identifier. To run optional AI features and verify premium status without an account, the app generates a random per-install identifier and uses Apple's App Attest to verify that requests come from a genuine copy of the app. These identifiers support subscription checks and abuse prevention; they are not advertising identifiers or used for cross-app tracking.
Subscription status. Premium subscriptions are processed by Apple and managed through RevenueCat. We receive whether you have an active subscription (keyed to the identifiers above, never your email); we do not receive or store your payment card details.
Usage and diagnostics. To understand which features are used and to find and fix crashes, the app sends product-analytics events to TelemetryDeck and crash/error diagnostics to Sentry. Product analytics include defined event names, counts, and feature details, such as the activity category used, completion of a breathing session, the number of mood labels selected, or a selected movement type. The app does not attach the text of your entries or conversations, your name, email, or recorded mood values to these analytics events. Diagnostics include technical error, app, and device information. Both services use pseudonymous identifiers, so these records should not be considered fully anonymous. Product analytics run only in released builds.
Service and security records. Our backend keeps operational records linked to your account identifier or device-attestation identifier, including AI request counts, subscription status, and request-verification information. We use these to enforce usage limits, check premium access, and prevent abuse. These are separate from aggregate AI token counts used to monitor cost.
We collect this data to provide and improve the app's features: saving and showing your entries, syncing them to your account, operating optional AI features and companion memory, unlocking premium, and maintaining reliability and security. We do not sell your data, and we do not use your health, fitness, or wellness data for advertising or marketing.
Where your data is stored and who processes it
- On your device. Your entries are stored on your device, so the app works instantly and offline. The app saves your ordinary AI conversation history on your device and syncs it to your account when you sign in. Temporary chats are held only for the current session and are not saved to history.
- Our backend (Supabase, hosting/database). If you sign in, your entries, saved AI conversations (including their context and favorites), and the companion's "memory" notes about you sync to your private account. Per-user access rules prevent other users from accessing this data, and our infrastructure provider encrypts it at rest. Authorized backend services and service providers process data as needed to operate Moonpool. If you don't sign in, your entries, saved conversations, and memory are not synced.
- AI providers (OpenAI and Anthropic). After you consent to AI features, the app sends the content a feature needs to our backend, which forwards it to either OpenAI or Anthropic to generate responses or update companion memory. Depending on the feature, that can include the reflection you are writing, recent wellness context, your conversation messages, the app's memory notes about you, and — if you've set them — your first name and preferred language. Wellness context can include information about mood, writing, gratitude, breathing, movement, and custom habits. In conversations, the AI can also look up your own synced entries to answer questions about your history. Our AI proxy does not store copies of request or response content after processing; saved entries, conversations, and companion memory are stored as described above. Operational records and aggregate cost metrics are described under "What we collect and why." Content sent through Anthropic's commercial API is not used to train its models under the standard terms we use. We do not use this content for advertising. Each provider's retention is described below.
- RevenueCat (subscription management) and Apple (Sign in with Apple, App Attest, in-app purchases).
- TelemetryDeck (product analytics) and Sentry (crash and error diagnostics). They process the usage events, technical diagnostics, and pseudonymous identifiers described above.
- Apple Health (optional). If you turn on the Apple Health toggles in Settings › Data, the app can write mood check-ins as State of Mind samples, completed breathing sessions as Mindful Minutes, and logged movement as workouts into Apple Health on your device. We do not read Health data back into the app. You can turn these off at any time; existing Health samples stay until you remove them in the Health app.
Each of these processors is required to protect your data consistent with this policy. We do not store personal health information in iCloud.
Reminders are local notifications scheduled entirely on your device — we run no push-notification servers. The Apple Watch app talks only to your iPhone over Apple's device-to-device connection; it makes no network connections of its own.
AI features and your consent
AI features are optional. Before the first time you use one, the app shows a disclosure and asks you to continue; nothing is sent for AI processing until you agree. After consent, AI processing includes responses you request and automatic maintenance of Luna's memory when remembering across chats is enabled and you send a message or request a reflection in a saved chat.
With Remember across chats on, sending a message or requesting a reflection in a saved conversation with Luna can trigger a memory update when there is new conversation history to process. Simply opening Luna, browsing saved conversations, or granting consent alone does not trigger AI processing. This can send prior conversation messages, existing memory notes, and relevant wellness context to either OpenAI or Anthropic to create updated memory notes. An update already started can finish after you close the conversation. Saved conversation history and the resulting memory notes sync to your account if you are signed in. You can stop using AI features at any time.
Remember across chats. You can turn this off in Settings → Companion. Luna then stops using saved memories, learning new memories, and automatically recalling earlier chats. Your current conversation and logged wellness activities remain available to AI features. Chats and saved memories are kept, and topics you have asked Luna to avoid still apply. Turning remembering back on restores saved memory without learning from chats written while it was off. Clear memory is a separate action. This choice syncs with your account; older app versions must update to continue using AI after it changes. Requests already sent for processing cannot be recalled.
Temporary chat. Choose a temporary chat when you do not want the conversation saved. Its transcript is held only for that session, does not sync, and does not create or update Luna's memory. Luna can still use your existing memory when remembering across chats is enabled, and your logged context, to respond. Temporary chat cannot save facts or change your profile, habits, reminders, or other settings; start a saved chat for those actions. Closing or leaving the temporary chat discards its transcript. Requests still pass through our AI service and either OpenAI or Anthropic, with the processing, service records, and provider-retention terms described in this policy.
Not medical care
Moonpool is a companion for reflection — not therapy, medical care, or a substitute for professional help. The app and its AI companion do not diagnose, treat, or give medical advice. If you are in crisis or thinking about harming yourself, please reach out to a person: someone you trust, or a free, confidential helpline near you via findahelpline.com (also linked in the app under Settings → About → Crisis resources).
Your choices and rights
- Export your wellness data as a JSON backup from Settings → Data. The file is created on your device; nothing is uploaded. This backup does not include AI conversations or companion memory.
- Optionally write to Apple Health from Settings → Data: mood check-ins as State of Mind, completed breathing as Mindful Minutes, and logged movement as workouts. Each toggle is off by default.
- Delete all your data from Settings → Data. This erases your entries, AI conversations, and companion memory from your device — and, if you're signed in, also erases the content of your synced entries, conversations, and memory from your account on all your devices.
- Sign out to stop syncing; or delete your account from the account screen, which removes your account, email, and all synced entries, conversations, and memory from our backend. Aggregate AI token counts and pseudonymous device-security records used for anti-fraud are not linked to your account and may persist.
- Manage or cancel your subscription through your Apple ID settings.
Depending on where you live, you may have additional rights (access, correction, deletion, portability). Contact us at the address above to exercise them.
Data retention
Data stored on your device is kept until you delete it or remove the app. Synced data is kept in your account until you delete the relevant entries or delete your account. When an entry deletion syncs to our backend, its content is erased; a small deletion record (the entry's identifier and timestamps, with no content) stays so the deletion reaches all your devices. Deleting your account removes those records too. Our AI proxy does not store copies of request or response content after processing; saved entries, conversations, and companion memory follow the storage and deletion rules above. Subscription-status records are kept while needed to operate the service.
OpenAI processing. OpenAI API content is not used for model training by default. We disable response storage; this does not remove standard abuse-monitoring logs, which may retain content for up to 30 days, or longer when legally required. See OpenAI data controls.
Anthropic's retention. Under the standard commercial API terms we use, Anthropic normally deletes inputs and outputs from its backend within 30 days of receiving or generating them. It may retain data longer to enforce its Usage Policy or comply with legal obligations. This retention is separate from model training: your API content is not used to train its models under these terms. See Anthropic's data-retention policy and model-training policy.
Security
Data in transit is protected with TLS. Synced data is isolated per account with row-level access rules and encrypted at rest by our infrastructure provider, requests are verified with App Attest, and sessions use standard token-based authentication with credentials kept in your device's Keychain.
Children
The app is not directed to children under 13, and we do not knowingly collect personal information from them.
Changes to this policy
We may update this policy as the app changes. Material changes will be reflected here with a new effective date.
Contact
Questions or requests: support@moonpool.me.