Text used with Voicepaper AI features is never used by Voicepaper to train AI models. This page explains how text is handled for AI Voice and AI Translation.
How Text Flows
For AI Voice, only the text needed to create audio is sent through Voicepaper's server to OpenAI / xAI or another selected provider. The generated audio is stored on your device and reused for repeat playback.
For AI Translation, the body, title, translation languages, and related information are sent through Voicepaper's server to OpenAI, and only the translation result is returned to the app.
How Voicepaper's Servers Handle Text
Voicepaper's servers relay AI Voice and AI Translation requests and manage usage credits.
They do not store the content of entered text, translated text, or generated audio.
Only management information such as request status, selected model or quality, and credit usage is recorded. The text body is not included in usage records.
For AI Translation, credits are committed only when the full translation succeeds. If the request is cancelled or fails, the reserved credits are released.
How AI Providers Handle Text
AI Voice uses OpenAI or xAI depending on the selected voice. AI Translation currently uses OpenAI.
OpenAI: Data sent to the API is not used to train or improve OpenAI models unless data sharing is explicitly enabled. Abuse monitoring logs may include input or output content and are retained for up to 30 days by default. Details: OpenAI data controls
xAI: xAI states that it does not train on API inputs or outputs without explicit permission. API requests and responses are temporarily stored for 30 days for abuse or misuse auditing, then deleted. Details: xAI API security FAQ
Related Articles
Related: Translate Text / Terms & Privacy