What leaves your Mac

This page is the full list, written plainly. The same facts in fewer words are on the privacy page.

Never

Your speech. Your transcripts. The audio of a meeting. Recognition, translation and style rewriting all run on your Mac, so there is nothing to send.

Only when you do something

Sign in, sign up, ask for a token — our own server. Sign-up carries your email, your password and a signature made with the install key.

Update check — verbaide.com/version.json, and only when you press the button in About. Never on its own.

Translation dictionary catalogue — cdn.verbaide.com, when Settings are open or when the language screen of the first-run setup is open.

Translation dictionary files — cdn.verbaide.com, when you download one.

Speech models — huggingface.co. The main one when it is first needed; a 98 MB term-dictionary model in the background on first dictation.

Style-model catalogue and manifest — cdn.verbaide.com/intelligence. Metadata only.

Style-model weights — huggingface.co/Qwen/…, when you start that download.

The dictionary and style-model requests carry no cookies and no authorization header.

Only with your consent

Usage counters. Numbers only — never text. The list: how many dictations and in which language or language pair; how long you spoke; how many characters came out (the count, not the characters); how long recognition, translation and styling each took; which paid feature was blocked; and your country, taken from the system region.

Two doors: by session when you are signed in, and when you are not, by a stable device id plus a signature made with the install key. If you did not allow statistics, none of this is sent.

The free tier

One download of the speech model. Without it even free dictation does not work. After that: no account and no limits.

Two requests are not tied to a plan at all and continue on the free tier: the translation-dictionary catalogue and the style-model catalogue, both fetched while Settings are open.