SpeakTrue

Privacy Policy for SpeakTrue

Last Updated: September 4, 2026

Overview

SpeakTrue is a speech app with text-to-speech, speech-to-text, speech-to-speech, soundboard management, and optional account sync. This policy explains what data is collected, how it is used, and how it aligns with mobile app store privacy requirements.

Scope

This policy applies to data handled by the SpeakTrue app, including data processed by service providers we use to operate core functionality.

Data We Collect

Below are the primary data categories relevant to app store privacy disclosures.

1) Identifiers and Account Data

• Account/session identifiers (for example: account ID) used for authentication and session management.
• If you sign in with Apple or Google, identity information provided by the auth provider and Supabase may include your name and email/account identifiers.

Purpose: App Functionality (sign-in, session restore, settings sync).

2) User Content

• Text you submit for hosted speech generation. This text is sent to ElevenLabs to create speech audio.
• Audio you record for transcription. These recordings are sent to ElevenLabs to create text transcripts.
• Audio you record for speech-to-speech. These recordings are sent to ElevenLabs for transcription, and the resulting text is then sent to ElevenLabs to generate speech audio.
• If you explicitly enable an experimental web transcription provider, the selected audio is instead sent to OpenRouter and its selected model provider, or to the configured Hugging Face inference service. For speech-to-speech, the resulting transcript is still sent to ElevenLabs for speech generation. These trials require separate provider-specific permission; the standard ElevenLabs agreement does not authorize them.
• Sample recordings you provide for hosted voice cloning. These samples are sent to ElevenLabs to create the cloned voice. SpeakTrue does not store those sample recordings after submission; they are transmitted for processing and deleted from SpeakTrue once the clone is completed or rejected.
• Reference audio, its matching transcript, and text you provide to the experimental Local Qwen Lab. These inputs remain on your device and are processed locally. SpeakTrue does not upload them to ElevenLabs, Supabase, or another speech provider from this workflow.
• Soundboard clip metadata and files when you choose cloud-backed soundboard features.

Purpose: App Functionality (generate speech, transcribe audio, manage clips).

3) Usage Data

• Limited app usage signals may be collected to support operation and service reliability.

Purpose: App Functionality.

4) App Settings and Preferences

• Voice/model selections, style controls, playback settings, theme, and related preferences.
• Settings may be stored locally and, when signed in, synced to your cloud profile.

Purpose: App Functionality and Product Personalization.

5) Diagnostics

• Crash data, performance data, and other diagnostic data may be collected to maintain app stability and improve quality.
• Diagnostic data is used for app functionality and analytics.
• Diagnostic data is not used for tracking.

Purpose: App Functionality and Analytics.

6) Optional Credentials (BYOK)

• Optional local fallback/provider keys that you provide in settings.
• Local fallback key is stored on-device and is not transmitted to SpeakTrue backend services.
• On iOS, new and successfully migrated fallback keys use device-only Keychain storage, available while the device is unlocked. A failed migration leaves the existing key in its previous storage and displays a storage error; plaintext removal is only claimed after successful migration.

Purpose: App Functionality (request routing and feature operation).

Data We Do Not Use for Tracking

• We do not use data for cross-app or cross-website tracking.
• We do not sell personal data.
• We do not share personal data with data brokers.
• We do not use your data for third-party advertising.

For app store privacy labeling, this means data is collected for app functionality/personalization, not tracking-based advertising.

Data Linked to You

Some cloud data is linked to your account when you sign in (for example: account ID, synced settings, and soundboard items). This linking is used to provide account-based features and continuity across sessions/devices.

Linked data categories include contact info (name/email), user content, user ID, and other usage data when collected for app functionality.

Diagnostic data (crash, performance, other diagnostic data) may be collected but is configured as not linked to your identity in app store privacy disclosures.

On-Device vs Cloud Processing

On-Device

• Local settings and app preferences.
• Temporary audio recordings created during recording/transcription flows.
• Optional local fallback key storage.
• Experimental Local Qwen Lab model weights, saved voice-reference profiles, and generated WAV files. Model weights are downloaded from Hugging Face; user reference audio, transcripts, and generation text are not sent with local inference requests.

Cloud (When Signed In or Feature Requires Backend)

• Authentication/session handling.
• Settings sync and soundboard metadata/storage.
• Speech processing requests routed through authenticated backend services.

Third-Party Services

ElevenLabs

Used for third-party AI speech processing. SpeakTrue sends only the content needed to provide the feature you request:

• Text-to-speech: the text you enter, plus speech settings such as voice, model, and playback parameters.
• Speech-to-text: the audio recording you choose to transcribe, plus transcription settings such as model, language, diarization, and audio event tagging.
• Speech-to-speech: your audio recording for transcription, plus the resulting transcript text for speech generation.
• Voice cloning: sample recordings of your voice or a voice you have permission to use. SpeakTrue does not store those voice-clone sample recordings after submission; they are transmitted to ElevenLabs for processing and deleted from SpeakTrue once the clone is completed or rejected.

SpeakTrue requires in-app permission before sending user content to ElevenLabs. ElevenLabs states its privacy and security practices in its own policy, including how it protects data it processes, and SpeakTrue relies on those practices to provide the same or equal protection for user content processed by ElevenLabs.

• Privacy policy: https://elevenlabs.io/privacy

Supabase

Used for authentication, database, edge functions, and optional storage.

• Privacy policy: https://supabase.com/privacy

Optional Web Transcription Trials

When an experimental provider is available to your account, its selector identifies OpenRouter or Hugging Face. A separate, initially unchecked permission control authorizes the selected provider to process the audio. Changing the model clears that permission. SpeakTrue validates the provider and agreement version before dispatching a request to that provider.

OpenRouter requests may be processed by the model provider selected through OpenRouter; Hugging Face requests use the configured inference endpoint. Provider-side retention and processing depend on that service and endpoint. SpeakTrue does not promise zero retention or on-device processing for these hosted trials. Do not submit sensitive or third-party recordings without appropriate permission. These hosted services are separate from downloading model weights for on-device Local Qwen inference.

Apple Services

Apple platform services may process app download, update, device, diagnostic, and account-related data according to Apple’s policies.

Microphone Access

• SpeakTrue requests microphone permission for recording/transcription, speech conversion, and on-device voice-reference profiles.
• Microphone is used only when you actively start recording.
• You can revoke access at any time in iOS Settings or Android system settings.

Data Retention

• Local data remains on your device until you delete it or uninstall the app.
• Cloud-stored soundboard content remains until you delete it.
• Temporary local audio files are deleted by app flows after processing.
• Voice-clone sample recordings are not retained by SpeakTrue after submission. They are deleted from SpeakTrue once the clone is completed or rejected.
• Local TTS voice-reference profiles contain recordings, transcripts and saved generation settings. New profiles are stored for the signed-in account on the device and hidden after sign-out; older unclaimed profiles remain shared within that app installation. Nothing uploads automatically. Profile files use iOS Complete file protection and are excluded from device backups. Downloaded models remain until you delete them or uninstall the app.
• Where optional account voice storage is enabled, Save to account uploads the selected reference recordings, transcripts and voice settings to private SpeakTrue storage for retrieval on other signed-in devices. This requires a separate confirmation and does not send those references to ElevenLabs. Model downloads and generated history are not included. Delete from account removes the cloud profile and queues recording deletion; previously issued download links can remain valid for up to five minutes, and storage cleanup may be retried. Independent and previously downloaded local copies remain until removed on their devices. Account deletion removes the account library and its cache on the requesting device; offline copies on other devices require local removal. Generated clip history remains installation-local and is not account-isolated.
• Account-linked settings remain while your account is active unless deleted/reset.

Security

• Data in transit is sent over HTTPS.
• Access to cloud-backed features is tied to authenticated sessions.
• Optional local fallback keys remain on-device.

Your Choices and Rights

You can:

• Review and modify settings in-app.
• Delete local app data by uninstalling the app.
• Revoke permissions (for example, microphone) in iOS Settings or Android system settings.
• Delete your signed-in account from the app through Settings → Account → Delete account.
• Request account and associated data deletion from the web at https://haddadios.github.io/SpeakTrue/account-deletion.
• Request account/data support through app store developer contact channels.

Children’s Privacy

SpeakTrue is not directed to children under 13, and we do not knowingly collect personal information from children under 13.

Changes to This Policy

We may update this policy as features or requirements change. The “Last Updated” date reflects the current version.

Contact

For privacy questions or requests, contact the developer through the app store listing contact methods.