Privacy Policy

1. Account information and Google sign-in

Authentication is handled by Clerk. Depending on how you register or sign in, account information can include an account identifier, email address, name, username, profile information supplied by the sign-in provider, and session information. Email/password registration is handled through Clerk; Google sign-in uses a Google identity credential to authenticate you through Clerk.

Google is used for account sign-in. The app does not request access to your Gmail messages, Google Drive files, or Google contacts. Google sign-in is not a connection to those Google services.

The service uses your authenticated account identifier to associate saved conversations, preferences, learning progress, and access permissions with your account. The app displays account details such as your name or email where available. Authentication providers process the information needed to manage sign-in, verification, and sessions.

2. AI conversations, preferences, and learning

When you submit a request, the service processes your prompt, relevant conversation context, and any supplied attachments to generate an answer. AI processing uses OpenAI through Replit's AI infrastructure. Selected prior conversation or learning information can be included as context; an individual request does not necessarily send every record on your account.

Signed-in conversations are saved on the service, including conversation identifiers, titles, timestamps, message text, and AI responses. Account-linked preferences can include your chosen companion, voice, speech settings, and other feature preferences.

Learning features store records such as lesson progress, submitted answers, attempts, feedback, and assessment results. Learning memory can save questions, answers, and learning feedback to support future responses. Its controls are separate from saved chat history.

When research is used, the service can send your request, relevant context, and derived search queries to OpenAI's search tools, and fetch selected public source pages. Information you put in a prompt may therefore be part of a research query. Avoid submitting passwords, authentication credentials, or information you do not want processed by these services.

Account-linked records are used to provide conversations, learning, personalization, and account access. Authorized service administration tools also access records for operating and reviewing the service. Saved account conversations are not automatically published as public website content.

3. Microphone access and voice processing

Voice features require microphone permission and an enabled voice feature. An active listening session may process microphone input continuously, including listening for an interruption while an answer is speaking or for a wake phrase when that mode is enabled. Android's floating or background voice features can continue an active session while the app is not the foreground screen.

Speech recognition

The web voice transcription path sends recorded speech audio to the service, which forwards it to ElevenLabs through Replit's integration infrastructure for transcription. The resulting transcript is used as your AI message and can be saved with the conversation. Browser or platform speech-recognition features, where used, are also subject to that browser or device provider's processing.

Native on-device recognition is used in supported native speech modes. Local recognition is distinct from the hosted web transcription path: even when recognition occurs on the device, the resulting text is sent to the service when you submit it for an AI answer. Processing can differ by app version, platform, selected mode, and availability.

Spoken answers

Local device speech can read answers without sending the text to a separate speech-generation service. For cloud-generated speech, the service sends the text to ElevenLabs. Generated audio can be stored in server object storage as a narration cache and downloaded or streamed to your device. Temporary device playback files and server-side cached audio have separate lifecycles.

You can stop listening using the voice controls and revoke microphone permission in your browser or device settings. Stopping listening does not delete transcripts, conversations, or audio already processed. This policy does not promise immediate deletion of audio by third-party speech providers.

4. Images, documents, screen captures, and sharing

If you choose an image or document, or approve a screen capture for a feature such as “Look at This,” the submitted content is processed by the service and can be supplied to the AI provider to answer your request. Uploaded files can be stored in server object storage. A one-time screen capture is processed as part of the submitted request; it should not be treated as content that stays only on your device.

Website-building features store drafts and selected media. If you publish a website, its published content and public website images are accessible to visitors. Public website media uses separate storage paths from private chat attachments. Choose carefully what you include in a published site.

If you use sharing controls, the selected response, text, link, or generated media is passed to your device's sharing system or the destination you choose. Information you share becomes subject to that recipient's or service's own handling.

5. Third-party services and information sent to them

The service uses the following providers for the features described here. These are functional service providers, not a statement that every provider receives every user's content.

• Clerk: account authentication, verification, and session management. Clerk Privacy Policy.
• Google: Google account authentication when you choose Google sign-in. Device or browser speech services may also involve Google depending on the platform. Google Privacy Policy.
• Replit: application hosting, database and object-storage infrastructure, and integration/AI infrastructure through which requests are processed. Replit Privacy Policy.
• OpenAI: AI answers, relevant prompt and conversation context, submitted content used for AI processing, and search/research requests. OpenAI Privacy Policy.
• ElevenLabs: speech audio submitted for transcription and text submitted for cloud speech generation. ElevenLabs Privacy Policy.
• Stripe: payment checkout and subscription/purchase processing when you use paid access. The app stores related customer, purchase, subscription, entitlement, and usage records. Payment details entered into Stripe's checkout are handled by Stripe. Stripe Privacy Policy.
• YouTube: embedded video content where that feature is used, including embeds using the youtube-nocookie.com domain. Loading or playing embedded media can connect your browser to YouTube; that domain does not mean no information is processed. Google / YouTube Privacy Policy.

Research can also connect to the public websites selected as sources. Links, embedded media, and destinations you open have their own policies.

Provider-side retention and handling are governed by the applicable service terms, configurations, and provider policies. This policy does not promise that providers immediately erase submitted data or make a blanket promise about their model-training practices. The disclosures above concern processing needed for implemented features; they are not a claim that submitted information is sold.

6. Browser/device storage and technical information

The app uses browser or device storage for preferences, lesson position, unfinished inputs, feature state, and authentication/session operation. Browser cookies can support sign-in and interface preferences. Installed web-app sharing can temporarily store shared text and links in a browser cache until the app consumes them; an abandoned share may remain until that cache is cleared.

The service processes technical information needed to serve and protect requests, including network/IP information used for rate limiting and diagnostic information such as request failures and provider error statuses. Hosting and external providers also process technical information under their own policies. This policy does not specify a fixed retention period for infrastructure logs or backups.

Clearing browser data or uninstalling an app affects local storage. It does not by itself delete server-side account records. Browser repair controls clear selected disposable app state rather than erase your account.

7. Retention, deletion, and your controls

Saved conversations, learning records, preferences, uploaded objects, and cached narration can persist beyond an individual session. Signing out does not erase saved conversations. There is no universal automatic expiry period stated here for these records, server audio caches, or provider records.

• Learning memory: the learning dashboard provides controls to disable future learning-memory capture, delete individual memory items, or clear that memory. These actions do not also delete separate chat, learning-progress, billing, or file records.
• Conversation reset: the conversation reset operation removes its saved message records but retains the conversation entry. It is not a complete account-erasure operation.
• Voice and local storage: stop voice features, revoke device/browser permissions, or clear local app storage using the relevant controls. These actions are distinct from deletion on the service.
• Other requests: contact freedeterminismai.support@gmail.com to ask about access, correction, or deletion of information associated with your account. Include enough information to identify the account, but do not send passwords or authentication tokens. Requests may require identity verification.

Deleting an item or an authentication account should not be assumed to erase all related database records, stored files, payment records, external-provider records, or backups. This policy does not promise a deletion deadline or removal from every provider or backup. We will explain what can be done for your request and any applicable limitations.

8. Access and security

Account-specific features use authentication and access checks. Public website publishing is a separate feature. No service can guarantee complete security, and information you intentionally publish or share is no longer limited to your account. Please avoid submitting information that is not necessary for the feature you are using.

9. Contact and policy updates

For privacy questions or requests concerning Ky El / Free Determinism AI, email the service operator at freedeterminismai.support@gmail.com.

This policy may be updated when the service's data practices change. The current version is available at this public URL, with its update date above.