The short version
Ghostly transcribes your voice on your device. Your audio never leaves your computer unless you
explicitly enable the optional AI Refinement feature, which sends only the transcribed text
(not audio) to a third-party language model provider you choose and configure with your own API key.
We do not collect analytics, telemetry, usage data, or personal information.
1. What Ghostly does locally
-
Voice transcription: Audio is captured from your microphone, processed by a
Whisper speech-to-text model, and converted to text — all on your device. Audio is never uploaded.
-
Transcription history: Ghostly stores a local history of your transcriptions in
your application data directory so you can review, copy, or delete past entries. This data never
leaves your device. You can clear it at any time from the History settings.
-
Settings and preferences: Your shortcuts, model choice, microphone selection, and
other settings are stored locally on your device.
2. Optional features that use the internet
Some features are off by default and only make network requests when you explicitly turn them on:
-
AI Refinement (Post-Processing): When enabled, the transcribed text is sent
to a language-model provider you choose (for example OpenAI, Anthropic, Apple Intelligence, or a
local model) to be cleaned up or rewritten. You supply your own API key. The request is made
directly from your device to that provider; Ghostly has no server in the middle, and we do not see
or store these requests. The provider's own privacy policy governs how they handle that data.
-
Model downloads: The first time you use Ghostly, it downloads a speech-to-text
model file from a public mirror. This download contains no personal data; it is the same public
file every user receives.
-
Update checks: If you leave "Check for updates" enabled, Ghostly periodically
contacts a version endpoint to see if a newer release is available. No personal information is
included in that request — only the request itself and your public IP address (as with any web
request).
3. What we do not collect
- We do not collect analytics, telemetry, crash reports, or usage metrics.
- We do not track you across sessions.
- We do not sell, rent, or share any data.
- We do not upload your audio, transcriptions, settings, or history to any server we control.
- Ghostly has no user accounts, no login, and no centralized servers that receive your data.
4. Permissions Ghostly requests on macOS
-
Microphone access: required to record audio for transcription. Recording only
happens while your configured shortcut is active.
-
Accessibility access: required so Ghostly can paste transcribed text into the
active application via a simulated keyboard event.
You can revoke either permission at any time in System Settings → Privacy & Security.
5. Your data, your device
Because all voice data and transcription history stay on your device, you control it entirely.
To export or delete your history, use the History settings inside Ghostly. To permanently remove
everything, uninstall the app and delete the Ghostly application data directory (accessible from
the About settings).
6. Children
Ghostly is not directed at children under 13 and does not knowingly collect any information from
anyone.
7. Changes to this policy
If this policy changes materially, we will update the "Last updated" date above and, where the
change affects how the app behaves, reflect it in the End User License Agreement shown when you
install or update Ghostly.