Effective 7 June 2026
Privacy Policy
Voiced Dictation is built for local speech transcription on your Mac. Audio and transcript text are not uploaded to Applification or to a cloud transcription service.
Local transcription
Voiced records only while push-to-talk is active. Temporary audio is used for local transcription, then deleted after processing, including error paths. Transcript text is copied or pasted according to your output setting and is not stored by Voiced after that flow.
Model downloads
On first run, Voiced asks you to choose a local transcription model and shows the approximate download size before downloading. Model downloads use the network and may contact Hugging Face through WhisperKit. Downloaded model files are stored in the app container and verified before loading.
Clipboard and Accessibility
Clipboard Only mode places the transcript on the clipboard for you to paste manually. Auto Paste mode temporarily writes the transcript to the clipboard and sends Command-V to the focused app after you grant macOS Accessibility permission. Voiced restores the previous clipboard contents when possible and does not collect clipboard contents.
Basic diagnostics
Voiced can share basic usage and crash diagnostics through PostHog to help us understand whether the app is working. Diagnostics can be turned off in Settings.
Included
- app opens and anonymous install activity
- app version, build number, macOS version, processor count, and architecture
- recording starts and cancellations
- transcription success or failure
- model load success or failure
- output failures and broad error categories
Not included
- audio recordings
- transcript text
- clipboard contents
- screen recordings or session replay
- file names, file paths, and window titles
- the apps or text fields you dictate into
Contact
For privacy questions, contact Applification through the support page at voiced.applification.net/support.