Privacy

Last updated: 9 August 2026 · The short version: we designed voiceflow so there is almost nothing to write on this page.

The application

voiceflow runs entirely on your machine. Your audio, transcriptions, history, vocabulary and configuration are stored locally and are never sent anywhere. The application makes exactly two kinds of network requests:

• Model download — once, from Hugging Face, when you install or change the speech model;
• Update check — an optional, anonymous request to the GitHub API at most once a day (disable with updates.check: false in the config).

There is no telemetry, no analytics and no account. The source code is public, so you can verify all of this.

This website

voiceflow.pbdevs.com uses a self-hosted instance of Umami to count visits. Umami is cookie-free and collects no personal data: no identifiers are stored, IP addresses are not retained, and nothing is shared with third parties. That is why you see no cookie banner — there is nothing to consent to. The site sets no cookies at all.

Contact

Questions? Open an issue at github.com/AveJaPl/voiceflow.