How Bulong handles your voice.
Bulong is private by architecture: your audio is transcribed on your Mac and never sent anywhere. Here is exactly what happens, in plain language.
Last updated: 8 July 2026
What is captured
Audio, and only while you are holding or have toggled recording on. It is written to a temporary file on your Mac and deleted right after it is transcribed. Nothing is captured when you are not dictating.
Where transcription happens
Entirely on your Mac. Speech is turned into text by NVIDIA Parakeet and Whisper running locally through CoreML. Your audio and the resulting text never touch a network.
What is stored
Two things, both on your Mac:
- Transcripts — kept in a plaintext file in Application Support that you can open, search, export, or delete at any time.
- Settings — stored in the standard macOS preferences for the app.
Nothing else. There is no cloud database and no hidden copy.
What leaves your Mac
Exactly one thing: downloading a speech model from Hugging Face the first time you use it. After the model is on your Mac, nothing leaves — no analytics, no crash reporting, no license check, no “phone home.”
How to verify it yourself
You do not have to take our word for it:
- Turn off your Wi-Fi and dictate. It still works.
- Or watch with Little Snitch or the macOS firewall — once the models are present, you will see no network traffic while you dictate.
Permissions and why
- Microphone — so Bulong can hear you while you dictate.
- Accessibility — so Bulong can type the result into the app you are working in.
Both are requested by macOS, shown in System Settings, and revocable at any time.
Contact
Questions about privacy? Email hello@bulong.app.