HeyYusuf
Privacy Policy
How HeyLanguages handles information for HeyYusuf.
Last updated:
HeyYusuf is a product of HeyLanguages. This policy explains how we handle information when you use HeyYusuf and its connected services. Contact us at dev@heylanguages.com or through HeyYusuf Support.
Information we process
- Account and profile information: your email address and authentication information when you register or sign in; your profile name, account identifier, selected dialect and learning level. Some guest AI features use an automatically created anonymous account identifier rather than an email account. An anonymous account is still an identifiable account within our service.
- Learning information: lesson and quiz progress, completion records, XP, saved practice state, learning preferences and related settings. Guest learning information may remain on your device; account features synchronize relevant information through our backend.
- Optional AI content: recordings submitted for pronunciation checking; Chat messages and conversation history; learning context; and saved conversation memory, which can include personal facts you share, words being learned and session summaries. Generated speech can contain information from a Chat reply.
- Subscription information: purchase and renewal records, store transaction information, subscription status and identifiers used to provide and restore Premium access. Apple or Google handles payment details for store purchases; we do not receive your full payment-card details from those purchases.
- Technical and security information: network/IP information, app and operating-system versions, device identifiers used by subscription services, request metadata and timestamps, status/error information and security or anti-abuse records. These support authentication, subscriptions, quotas, troubleshooting and service protection.
- Support information: your contact details and the content you choose to send when requesting help or exercising privacy rights.
The app also stores settings, consent choices, cached audio and saved learning state on your device. Information used only on your device is distinct from information transmitted to our service providers. Avoid including sensitive personal information or information about other people in practice messages, recordings or support attachments unless necessary.
How information is collected
Some information comes directly from you, such as account details, submitted recordings, Chat messages and support correspondence. Other information is processed automatically through network requests and integrated services; you do not need to fill in a form for this to happen.
For example, opening the app, restoring or refreshing a session, synchronizing account progress, checking subscription access and downloading media can contact our backend or the relevant service. These requests expose connection information such as an IP address to the receiving service and can include app/device information, account or service identifiers, and request headers. Backend services retain operational and security records such as request timestamps, status codes and errors. RevenueCat's subscription integration automatically sends app/user information and an iOS vendor identifier. When an authentication flow invokes Cloudflare Turnstile, its integrated verification service processes connection and device/browser signals and challenge results.
These necessary non-AI requests are separate from permission to send content to AI providers. Declining AI sharing blocks the optional AI transmissions described below, not ordinary authentication, subscription or security processing. Support opens your email application; HeyYusuf supplies a subject but does not automatically attach recordings or diagnostic logs. You choose the message and any attachments before sending.
How we use information
We use information to provide learning activities, save and restore progress, personalize relevant practice, operate optional AI features, administer purchases and subscriptions, provide support, detect abuse, troubleshoot problems and comply with legal obligations. RevenueCat also processes subscription information for subscription reporting and analytics.
Where applicable law requires a legal basis, we rely on providing the services you request, consent for optional AI sharing, legitimate interests in security and service operation, and legal obligations. You can withdraw AI-sharing permission as described below without losing access to ordinary non-AI learning. Existing subscription, connectivity and access requirements still apply.
Optional AI features: recipients and data
AI requests pass through our Supabase backend. Supabase provides authentication, database storage and server functions, including account-linked conversations, memory, learning records and access controls.
| Recipient | Information sent and purpose |
|---|---|
| OpenAI | The pronunciation recording you submit is sent for Arabic transcription. HeyYusuf uses the returned transcript to evaluate pronunciation. The expected phrase, dialect, activity context and any evaluation hint are also sent to our backend; these are separate from the audio-transcription request made to OpenAI. |
| Anthropic | Chat messages, relevant conversation history and context are processed to generate replies. Depending on the feature, context includes your displayed name, learning level, dialect, learning material and saved memory. Conversation text and existing memory may also be processed to produce memory summaries for later practice. |
| ElevenLabs | Text is sent to generate spoken audio using a voice selected for your dialect. This can include personal information present in a Chat reply. This speech-generation flow does not send your microphone recording to ElevenLabs. |
Playing an existing packaged or cached recording does not itself request new AI speech generation. Downloading existing media may still contact its hosting service.
Your AI-sharing choice
In iOS versions that include Profile → AI privacy, HeyYusuf asks for explicit permission before sending information to these AI providers. The prompt names the recipients and information shared. Choose Allow to enable optional AI processing or Not now to decline. Microphone permission is a separate device permission; granting it does not by itself grant AI-sharing permission.
With AI sharing off, you can continue ordinary non-AI lessons and existing recorded listening. Speaking activities provide an unassessed continuation where applicable; they do not claim a successful pronunciation result. AI Chat and new AI-generated audio require permission. While permission is on, related automatic operations, such as conversation-memory summarization, can occur without a separate prompt for each request.
To withdraw, open Profile → AI privacy → Turn off AI sharing. The choice is stored for the account or guest context on that installation. It is not a server-wide setting and does not automatically change another device. If the app reports that the choice could not be saved, retry before closing the app; requests remain blocked in the current session.
Withdrawal stops new AI requests from that installation. It cannot recall a request already sent or prevent that request from finishing processing or saving its result. It does not delete existing conversations, memory or provider records. Account deletion and other deletion requests are explained below.
Older app versions may not include these controls. Updating this policy does not add controls to an installed app. If your version does not show AI privacy, avoid AI features until an updated version with these controls is available, or contact us for help.
AI-provider training settings and retention
Our current provider settings are: OpenAI voluntary data sharing disabled; Anthropic standard retention with no participation in its training program; ElevenLabs model-training sharing disabled. These settings are distinct from retention, safety processing and deletion, and do not reverse any earlier processing.
- OpenAI: We use the audio-transcription API. OpenAI excludes API content from model training unless sharing is enabled. Its current endpoint documentation lists no abuse-monitoring or application-state retention for audio transcription. This is not a promise that all technical metadata is immediately erased or that HeyYusuf has an account-wide zero-retention agreement. OpenAI data controls.
- Anthropic: Under standard commercial API retention, inputs and outputs are normally deleted within 30 days. Safety enforcement, legal obligations and contractual exceptions can require longer retention. Commercial content is not used for training by default; explicit opt-ins and deliberately submitted feedback have separate rules. Retention policy and training policy.
- ElevenLabs: Disabling model-training sharing excludes newly submitted data from training according to its terms, but does not disable generation history or erase earlier data. Generated text and audio can remain in history. History deletion can leave debugging/moderation records and backups; its documentation describes backups lasting up to 30 days after deletion. We do not claim a fixed automatic expiry for all TTS history or that Zero Retention Mode is enabled. Training controls and retention/deletion guidance.
Provider processing is subject to the applicable service terms and legal requirements. Turning off sharing in HeyYusuf does not automatically send deletion requests to these providers.
Other service providers and subscriptions
We use RevenueCat to validate, manage and restore Premium access. It receives subscription/transaction information, app-user identifiers and device/app information, including an iOS vendor identifier. Apple and, where available, Google process store purchases under their own policies.
Cloudflare Turnstile protects relevant authentication requests against abuse. Its verification flow can process IP address, browser/device signals, connection characteristics and challenge results. Cloudflare processes signals on our behalf for site protection and separately acts as a controller when using signals to improve bot detection. This security processing is separate from optional AI sharing. See Cloudflare's Turnstile Privacy Addendum and Privacy Policy for these distinct roles.
Our iOS app does not use these AI features for advertising or cross-company advertising tracking. We do not infer that giving AI permission also gives advertising permission. Other platform versions may offer different features; the iOS controls described here do not change those versions.
Third-party privacy safeguards
HeyYusuf remains responsible for the third-party sharing it enables. Apple's privacy rules require recipients to provide protection at least equivalent to this policy and its applicable privacy requirements; outsourcing processing does not remove that responsibility.
The published data-processing terms of OpenAI, Anthropic, ElevenLabs, Supabase, RevenueCat and Cloudflare describe safeguards for the processing they cover. These include documented processing instructions, security and access controls, assistance with applicable privacy rights, and safeguards for subprocessors. The applicable terms, legal scope and provider roles differ; Cloudflare's separate controller processing is described above. Store-operated payment processing is also subject to the relevant store's policies.
These safeguards do not mean that every provider has identical terms, that all data is erased immediately, or that no legally required or security-related retention occurs. The provider-specific settings, retention limitations and deletion routes in this policy still apply.
Retention and account deletion
We retain account, progress, conversation and memory records while needed to provide the relevant service, or until deleted through the applicable process. Security records, support correspondence, transaction records and backups may have different retention periods where needed for security, legal obligations, disputes or operational recovery. We do not promise immediate deletion from every backup or third-party system.
In iOS versions containing the updated deletion controls, open Profile → Delete Account to initiate deletion of your signed-in account. If you have an existing anonymous AI identity, Profile instead offers Delete Guest AI Account. Both require confirmation and an online connection. A guest using only local lessons may have no server account to delete.
The service verifies the requesting identity and deletes its associated active HeyYusuf application records and authentication account. Deleting an isolated guest AI account turns off guest AI sharing and preserves device-only lesson progress. It does not cancel purchases or erase records a store or provider legitimately retains. Local cached files and information on another installation are not a promise of simultaneous remote erasure.
Deleting your HeyYusuf account does not cancel an Apple App Store or Google Play subscription. Manage or cancel subscriptions in the relevant store settings.
For access, correction, deletion or questions about information previously processed by a provider, contact dev@heylanguages.com. We may need to verify your identity and coordinate with the relevant provider. Our current in-app deletion process does not automatically identify and delete every provider's generation-history item. We will explain applicable limitations and lawful retention when handling a request. You can also find deletion information on our Delete Account page.
Your rights, security and international processing
Depending on where you live, you may have rights to access, correct, delete or receive your information, restrict or object to processing, withdraw consent, and complain to a relevant privacy authority. Contact us to exercise those rights. Withdrawal does not undo processing that already occurred.
We use access controls and secure connections appropriate to the service. No system can guarantee absolute security. Our providers may process information outside your country, subject to their applicable contracts and safeguards.
Children and policy changes
HeyYusuf is not intended for children under 13. If you believe a child under 13 has provided personal information, contact us so we can investigate and address it appropriately.
We may update this policy as our services or practices change. The date above identifies the latest revision. Contact: dev@heylanguages.com.
Related pages
Questions? Contact dev@heylanguages.com.