PodTalk Privacy Policy

1. Who is responsible

PodTalk is published and operated by Ahmet Yasir Demir. Ghaith Labs is the brand used to present PodTalk and its policy pages. This policy covers the PodTalk app and its supporting services. For privacy requests, contact dydemiryasir@gmail.com.

2. Information we process

• Account and profile: an automatically generated Firebase account identifier, username, display name, profile information and social activity such as follows and likes. You can use an account without giving us an email address or password.
• Source material: the topics, story ideas, pasted text, documents and images you choose to submit, including text extracted from selected documents or images.
• Generated content: scripts, summaries, episode metadata, cover images, generated speech and quiz questions, together with your language, voice, speaker and duration selections.
• Activity and entitlement records: generation requests, processing outcomes, allowance counters, subscription status, playback and feature-use events, and technical information produced by the app and its service providers.
• Community and support records: comments, content and user reports, block records, feedback messages and related app version, platform and language information. If you email us, we also receive your email address and the information you include.
• Device information: local preferences, playback position and cached audio. When notifications are enabled, Firebase Cloud Messaging uses a device registration token to deliver them.

Source material can contain personal information about you or other people. Submit only material you have the right to share, and remove confidential information that is not necessary for the requested output.

3. Device permissions and selected-content access

PodTalk requests access in connection with the feature you choose to use. A device permission authorises access controlled by your operating system; it does not, by itself, authorise sharing that material with an AI provider or publishing it to other users. AI processing is subject to the separate disclosure and permission described below. System prompts and available choices depend on your device, operating-system version and previous permission decisions.

Camera and document scanning

Opening the document-scanning screen starts the camera and may trigger the camera permission request. Access is used to photograph the document you select. The app uses on-device text recognition to extract text from captured images, creates a temporary text file and passes that text to the material-import screen. When you submit it for generation, the extracted text is uploaded and processed as described in the AI section. Granting camera access does not itself start an AI request. If access is refused, camera scanning is unavailable; you can instead type or paste text, or import a supported file.

Photo selection and saving generated videos

The gallery action on the scanning screen opens a system image picker. Images you select are used for on-device text recognition, followed by the same material-import process. A system selection interface may provide access to selected items without granting access to the entire photo library. If you cancel the selection or access is unavailable, no images from that selection are added.

When you choose the podcast-video export feature, the app combines episode artwork and audio into a video and attempts to save it to your photo library for sharing, including Instagram-style sharing. Your device may request photo-library access for that save operation. If saving is refused or fails, that gallery export cannot complete; generating an episode does not require this permission. Saving a video does not automatically post it to a social network. Copies saved in your gallery are controlled separately from records in PodTalk.

Documents and other selected files

Material import opens a system file picker for supported PDF, DOCX, TXT, PNG and JPEG files. This gives the app access to the files you select rather than unrestricted access to your documents. Depending on the format and available extraction, the app submits extracted text or the selected file when you request generation. If you cancel, that selection is not imported. You can use a topic or pasted text instead.

Notifications

Before the first generation for which notification permission has not already been addressed, the app can offer an optional notification explanation. If you choose to receive notifications, it requests operating-system permission for alerts, sounds and badges. You can also request permission through the notification switch in Settings. Firebase Cloud Messaging uses a device registration token, associated with your app account, and your preferred language to deliver notices such as generation completion. Tapping a generation notice opens the relevant episode or story. Refusing notifications does not prevent generation; check your library for the result.

To turn notifications off, use your device’s settings; the in-app switch opens those settings when needed. Disabling delivery does not itself delete a registration token already stored in our database. Permission changes and data deletion are separate actions.

Audio playback, tracking and permission controls

The current creation and scanning features do not record your microphone. Synthetic narration and background playback use generated audio. Background playback and network access support listening and online features; they are not a request to record surrounding sound.

The updated app removes the Meta/Facebook SDK and the iOS App Tracking Transparency request previously included in earlier releases. This change does not remove Firebase Analytics or the application event records described in the infrastructure section, and does not erase information previously processed by an earlier release.

You can change available camera, photo-library and notification permissions in your device’s app settings. Withdrawing device access does not recall material already submitted, remove imported copies or delete exported files. Use the account-deletion process or contact us for a request concerning information already held by PodTalk. Cancelling a picker or refusing one of these optional device permissions affects the associated feature; new AI processing still requires the separate AI permission.

4. AI processing and your permission

In the updated app, selecting Create or turning on Settings → AI data sharing opens the same disclosure if permission has not already been granted. The Settings switch starts off when no current permission is stored and turns on only after you select “Allow and continue” and the choice is saved on that device. Closing the disclosure leaves it off. A saved grant applies to both entry points; the app does not request the same permission for each new episode.

This permission is needed because generating a new script, quiz, cover or uncached synthetic speech requires sending the relevant material to the providers identified here. If you close the disclosure without agreeing, the app does not start new AI operations and those features remain unavailable until you grant permission through Create. Device permissions, notification permission and a subscription do not substitute for this choice.

• Google Gemini: receives your topic, story idea, pasted text, uploaded material or extracted text to generate scripts and summaries. Lesson text is also sent when generating a quiz.
• ElevenLabs or Inworld: receives the script segments and voice/language selections needed to generate speech. Which service receives a segment depends on the selected episode voice.
• Kie.ai: receives a cover-image prompt derived from the generated content and returns an image for the episode.

You can withdraw permission by turning off Settings → AI data sharing. This prevents new AI operations from that app installation and stops playback. Requests already submitted can finish, and withdrawal does not recall information already sent. To grant permission again, turn that switch on or select Create and accept the disclosure. Permission is stored locally; it is not a separate server consent record.

Provider processing is subject to the applicable service terms and configuration. We do not promise that every provider uses the same retention period or model-improvement rules. In particular, Gemini distinguishes paid and unpaid services in its API terms. Do not submit sensitive or confidential material on the assumption that all AI processing has zero retention.

Earlier app releases can differ from the updated permission and account-management flows described here. Update the app to use those controls.

5. Infrastructure, purchases and analytics

Google Firebase provides account authentication, Firestore database storage, file storage, server functions, messaging and remote configuration. Firebase Analytics and application event records help us understand usage, diagnose failures and operate generation allowances. The app’s custom event handling is designed to record product usage rather than source material and filters text, prompt, email and source-URL fields. Provider and infrastructure records can still contain request-related information and account identifiers.

RevenueCat verifies subscription entitlements using the app account identifier and store purchase information. App Store purchases and billing are handled by Apple; the app does not collect your payment-card details. Store and service providers also process technical information under their own terms.

Relevant notices: Firebase privacy and security, Google Privacy Policy, RevenueCat Privacy Policy, ElevenLabs Privacy Policy, Inworld Privacy Policy, Kie.ai Privacy Policy and Apple Privacy Policy.

These are the providers’ public notices. API processing can also be governed by separate service or business terms; a provider’s website privacy notice alone does not establish how every API input is retained or used. Direct questions about PodTalk’s handling of your information to dydemiryasir@gmail.com.

6. Why information is processed

We process information to create and deliver requested episodes, speech, covers and quizzes; maintain your library and profile; verify paid access and enforce allowances; support community interactions; respond to feedback and reports; diagnose errors; and prevent misuse. Where applicable law requires a legal basis, we rely on your permission for the disclosed AI sharing, performance of the service you request for account functions, legitimate interests in operating and securing the service where those interests are not overridden by your rights, and compliance with applicable legal obligations. Permission to use AI is not permission to publish your source material publicly.

7. Profiles, public content and sharing

Profile details and social activity are visible through the app’s community features. Publishing to the public episode feed is controlled by moderation privileges in the current implementation; ordinary generation requests are kept out of that feed. Public episodes and comments can be viewed by other signed-in users.

An exported file, shared text or a link can be copied by its recipient. Media links can continue to work while the underlying file or access token remains available. Deleting a record in the app cannot remove copies that another person has saved.

8. Storage and deletion

Account records, source uploads, generated content and operational records are stored to provide the app and its library. Source uploads and server audio caches do not have one universal automatic expiry in the current implementation. Local preferences and audio caches remain on your device until cleared by the app, account-deletion flow or device controls.

To delete your account in the updated app, open Settings → Delete my account and confirm. You are signed out, and a background process removes the Firebase account and linked application records handled by the deletion workflow, including profile/library records, public copies, comments, feedback, reports, blocks, username reservations, generation usage, product events, uploaded materials and identifiable episode assets/cache entries.

Cleanup runs in scheduled intervals and normally completes approximately 15–30 minutes after the request. Failed attempts remain queued for retry. This is an expected operating time, not a guaranteed deadline. If deletion remains unresolved after 24 hours, contact dydemiryasir@gmail.com. For an earlier app version or if you cannot access the app, request deletion by email and include your account identifier if available. We may need information to establish that the request concerns your account.

Removing an individual episode is not the same as deleting an account and does not guarantee the immediate removal of every related cached file. Account deletion does not cancel an Apple subscription, erase another person’s exported copy, or guarantee immediate erasure of independent provider records, store transaction records, security logs or infrastructure backups. These records have separate service or legal retention requirements.

9. Your choices and privacy requests

You can manage AI permission and account deletion in Settings, manage available device permissions as described in the permissions section, and manage Apple subscriptions through your Apple Account. Depending on applicable law, you can request access, correction, deletion or other rights concerning your personal information, and raise a concern with your local data-protection authority. Send requests to dydemiryasir@gmail.com; do not send account passwords or full payment-card details.

10. Children and age eligibility

The baby and child story settings describe the audience of a story; they are not account-age verification or parental controls. AI providers have their own age requirements. Gemini’s API terms currently restrict use in clients directed to, or likely to be accessed by, people under 18. Adults should handle submissions and should not include a child’s name, image, school, contact details or other identifying information in a prompt or document. Contact dydemiryasir@gmail.com if you believe a child’s personal information has been submitted.

11. Processing locations, security and updates

Firebase and AI services process data through their infrastructure and subprocessors, including across borders. This policy does not promise storage exclusively in a particular country. HTTPS connections, Firebase authentication and application access rules are part of the implementation. AI processing is not end-to-end encrypted against the services that process the content. No service can guarantee complete security.

We update this policy when the service or its processing changes and identify the revision date above. A material change to AI recipients or the information shared requires an updated disclosure and renewed permission in the app. Related information appears in our Security page and Terms of Use.