Privacy, license and credits

Last updated 13 October 2026

Privacy

The Cadence app

• Your speech is turned into text on your Mac. Cadence has no account, no server and no networking code, so nothing you say is sent anywhere.
• It uses the microphone only while you hold a dictation key, and Accessibility only to put text where your cursor is.
• Dictation history and settings are stored on your Mac in ~/.cadence, in an encrypted database. Audio is saved only if you turn that on, and you choose how long it's kept.
• Delete everything by quitting Cadence and removing ~/.cadence and the app.

This website

• No cookies, analytics, trackers, forms or third-party scripts. The page's security policy blocks them outright.
• The site is hosted on Cloudflare Pages, and downloads come from GitHub. Like any web host, both handle standard request data (such as your IP address) to serve and protect them.

License

Cadence is open source under the MIT license. You may use, copy, modify and share it, including commercially, as long as the copyright notice comes with it. It is provided "as is", without warranty of any kind. The full text is in the LICENSE file.

Credits

Cadence stands on these open-source projects. Their full license texts ship inside every download as THIRD_PARTY_NOTICES.md.

• Whisper speech-recognition model by OpenAI (MIT).
• whisper.cpp by Georgi Gerganov and contributors (MIT), used through whisper-rs (Unlicense).
• SQLCipher by Zetetic (BSD-style), through rusqlite (MIT), with OpenSSL (Apache 2.0).

Designed and built by Daniel Cuesta under DC Labs, with Claude.

Trademarks

Wispr Flow, Superwhisper, Apple, macOS and OpenAI belong to their owners. They're named here only to compare features and to credit the work Cadence uses. Cadence is an independent project and is not affiliated with, sponsored by or endorsed by any of them.

Contact

Questions and bugs: open an issue on GitHub. For a security problem, use Report a vulnerability on the repository's Security tab instead, so the details stay private.