العربية
تخص هذه الصفحة تطبيق رفيق للقراءة العربية من بيان للذكاء الاصطناعي، وليست سياسة عامة لجميع خدمات الأكاديمية.
التعلّم على الجهاز
لا يطلب رفيق حسابًا أو اسم المتعلّم. تبقى ملاحظات مراجعة الكلمات في ذاكرة الجلسة ولا تُحفظ في سجل تقدم دائم. القصص والسرد الصوتي المضمّن يعملان دون رفع بيانات المتعلّم. القراءة اليدوية لا تتطلب تسجيل الصوت. تدريب النطق الاختياري له موافقة منفصلة قبل التسجيل والإرسال.
أدوات تسجيل البالغين الاختيارية
بعد إذن الميكروفون، يمكن للبالغ تسجيل صوته للاستماع إليه. يُخزّن التسجيل مؤقتًا على الجهاز. اختيار مسح الجلسة أو قصة أخرى يحذف تسجيلات الجلسة؛ قد تبقى ملفات مؤقتة عند الإغلاق المفاجئ. يمكن إزالة البيانات المحلية بحذف التطبيق.
عند اختيار التفريغ وتأكيد أن التسجيل يخص بالغًا موافقًا، يُرسل الصوت عبر اتصال مشفّر إلى خادم رفيق على medresearch-academy.om ثم إلى OpenAI لتحويله إلى نص. لا تستخدم هذه الميزة لتسجيل طفل أو معلومات حساسة. يعيد الخادم النص للتطبيق ولا ينشئ سجلًا دائمًا للصوت أو النص في قاعدة بيانات رفيق. قد تعالج خدمات الاستضافة والمزوّد سجلات تشغيل وأمان تشمل عنوان الشبكة IP ووقت الطلب وحالته وبيانات تقنية، وفق سياساتها، لتشغيل الخدمة ومنع إساءة الاستخدام، وليس للتتبع الإعلاني. تذكر وثائق OpenAI الحالية أن نقطة التفريغ الصوتي لا تحتفظ بمحتوى الصوت في سجلات مراقبة الإساءة أو حالة التطبيق؛ يُعالج الصوت لإرجاع التفريغ المطلوب. لا يعني حذف الصوت من الجهاز حذف أي سجلات يحتفظ بها مزوّد الخدمة.
تدريب النطق الاختياري
في الإصدارات التي تدعم الميزة، يوافق ولي الأمر أو بالغ مخول منه قبل تسجيل وإرسال مقطع لا يتجاوز ٢٩ ثانية والنص المعروض عبر خادم رفيق إلى Microsoft Azure Speech لتقييم النطق. يرسل الصوت عند إنهاء المحاولة أو بلوغ الحد الزمني. لا نطلب اسم الطفل أو نستخدم صوته للتعرف على هويته. النتائج اقتراحات تدريب قد تخطئ، وليست درجات مدرسية أو تشخيصا. لا نحفظ التسجيلات أو النتائج في قاعدة بيانات. يحذف الملف المؤقت بعد المحاولة أو الإلغاء؛ قد يبقى عند الإغلاق المفاجئ ويمكن إزالته بحذف التطبيق. توضح Microsoft أنها لا تحتفظ بالصوت أو النص في تقييم النطق. قد تبقى سجلات تقنية للشبكة. يمكن سحب الموافقة من شاشة التدريب؛ لا يمكن إلغاء معالجة اكتملت. تظل القراءة والاستماع متاحين دون مشاركة الصوت.
المزوّدون والاختيارات
الاستضافة لدى Vercel والتفريغ لدى OpenAI. أُنتج السرد المضمّن عبر ElevenLabs؛ لا يُرسل تسجيل المستخدم إلى ElevenLabs. لا يتضمن التطبيق إعلانات أو تتبعًا إعلانيًا أو مشتريات. يمكنك رفض إذن الميكروفون واستخدام القراءة والمراجعة اليدوية دون التفريغ السحابي.
التواصل
لطلب المساعدة أو الاستفسار عن البيانات، راسل info@medresearch-academy.om. إذا راسلتنا، نتلقى عنوان بريدك ومحتوى رسالتك للرد على الطلب. لا تُرسل بيانات طفل أو تسجيله في طلب الدعم. تتطلب طلبات البيانات معلومات كافية لتحديد الطلب؛ لا يتيح التطبيق حسابات يمكن البحث عنها بالاسم.
English
This notice covers the Rafiq Arabic reading app operated by BAYAN AI TECHNOLOGIES LLC.
No learner account or name is required. Word-review observations remain in session memory, without a permanent progress record. Manual learner practice does not record or upload audio. Versions with optional pronunciation practice require separate consent before recording and uploading. Bundled stories and narration can be used offline.
Optional adult recording requires microphone permission. Audio is stored temporarily in device cache. Clear the session or choose a new story to remove session recordings; temporary files may remain after an unexpected closure. Uninstalling the app removes its local data.
When an adult explicitly requests transcription and confirms consent for their own adult recording, audio is sent over HTTPS to our server at medresearch-academy.om and then to OpenAI. Do not upload child recordings or sensitive information. Our transcription handler returns text and does not create a permanent audio or transcript database. Hosting and processing providers may handle operational or security logs, including IP addresses, request times, response status and technical diagnostics, under their policies. These logs support service operation and abuse prevention, not advertising tracking. OpenAI currently lists no abuse-monitoring or application-state retention for its audio transcription endpoint; audio is processed to return the requested transcript. Deleting local audio does not erase provider-held logs.
Optional pronunciation practice
In versions offering this feature, a guardian or authorized adult must consent before recording and sending up to 29 seconds of reading audio and the displayed text through our server to Microsoft Azure Speech. Audio is sent on finishing or reaching the time limit. No child name is requested and voice is not used for identity recognition. Feedback is a practice suggestion, may be wrong, and is not a school grade or diagnosis. We do not create an audio or results database. The temporary attempt file is deleted after assessment or cancellation; unexpected closure may leave cache files, removable by uninstalling. Microsoft documents no audio or text retention for pronunciation assessment. Hosting may process network and security logs. Consent can be withdrawn from the practice screen; completed processing cannot be undone. Listening and manual practice remain available without audio sharing.
Microsoft Speech data processing
Vercel hosts the server and OpenAI supplies transcription. ElevenLabs generated the bundled narration; user audio is not sent to ElevenLabs. The app contains no advertising, advertising tracking or in-app purchases. Recording and cloud transcription are optional; listening and manual review remain available without them.
Contact info@medresearch-academy.om for privacy questions or data requests. Support emails provide us your email address and message for responding to your request. Do not include a child’s information or recordings. We need enough information to identify a request; there is no named learner account to search.
Provider information: OpenAI privacy · Vercel privacy.