What Happens When You Upload Audio
When you upload a file to a cloud transcription service, it is transferred to the provider’s servers, processed by their speech recognition system, and the transcript is stored in your account. Depending on the provider, the audio and text may be kept until you delete them, retained for a period afterwards in backups, processed by sub-processors, or — under some terms — used to improve their models.
None of this is necessarily malicious. It is simply how cloud software works. But it means the privacy of your recording depends on the provider’s policies, security and future decisions.
Questions to Ask Any Transcription Provider
Before uploading sensitive audio, find clear answers to these questions in the provider’s privacy policy and terms:
- How long are audio files and transcripts retained, including backups?
- Can my content be used to train or improve AI models? Is that opt-in or opt-out?
- Do employees or contractors ever listen to recordings for quality review?
- Which sub-processors handle my data, and in which countries?
- Can I permanently delete my files and account?
Who Should Avoid Cloud Transcription
For a podcast episode that will be public anyway, cloud transcription may be fine. For many other recordings, it is worth thinking twice: journalists protecting sources, researchers bound by ethics approvals, lawyers handling privileged conversations, healthcare and therapy professionals, HR teams, and anyone recording family or personal matters.
The On-Device Alternative
Modern phones and laptops are powerful enough to run speech recognition locally. With on-device transcription, the recording is processed where it was made, and there is no provider copy to worry about.
VoiceMaster is built entirely around this model: the speech model ships with the app, there is no account and no cloud sync, and recordings and transcripts stay in your local library on iPhone, iPad or Mac. You decide if and when anything is exported.