All help topics

What Data Is Collected

What stays on your Mac, what we never receive, and what the app does collect.

EnviousWispr collects anonymous usage data and crash reports. Nothing you say is part of that. Your audio and your transcripts never reach Envious Labs, the company that makes the app.

What stays on your Mac

Your transcripts are saved to your History so you can find them later. They sit in your user folder, and Envious Labs never receives a copy.

If you turn on Escape Recovery, a recording you cancel with your keybind is transcribed and held in that same folder, so you can paste it back or press Keep to make it permanent. It stays available for 24 hours. After that it is removed while the app is running, or the next time you launch it. Its audio is deleted once its text is safely saved, exactly as with any other dictation.

While you dictate, the app keeps an encrypted backup of the recording. Once your text is safely saved, the app requests deletion of that backup. If the app quits first, EnviousWispr makes one recovery attempt the next time it runs, then requests deletion whether that recovery succeeded or failed.

Your custom words, settings, and API keys live here too. They stay on your Mac unless you turn on cloud polish, which is covered below.

What Envious Labs never receives

  • Your audio
  • Your transcripts, before or after polish
  • Your custom words
  • The text around your cursor
  • Your API keys
  • Your name or email address

What the app does collect

The app collects anonymous usage and crash data. That data shows whether a release broke dictation on a particular macOS version, or whether anyone ever opens a setting that took a month to build. It is on by default and cannot be turned off.

The app records how you use it, never what you said. There is no account, and nothing in the data names you, although each installation gets a random ID so that one Mac counts as one user. The privacy policy has the full detail.

If you would rather run without it, EnviousWispr is open source under the GPLv3 license. You can build the app yourself, and dictation works in exactly the same way.

Where your text goes if you use cloud AI polish

Polish runs on your Mac by default. If you choose OpenAI, Gemini, or Claude instead, you add your own API key. Your transcript is then sent to that provider, along with your custom words and the name of the app you are dictating into, so the model gets your spellings and tone right. Audio is never sent. The app tells you this when you set it up. That connection is your account with that company, governed by their terms.

Ollama works in two ways, and only one of them keeps your transcript on your Mac. A model you download runs on your Mac and sends nothing anywhere. A hosted model runs on Ollama’s servers, so your transcript goes to Ollama in the same way it would go to any other cloud provider. EnviousWispr lists the two kinds under separate headings so you can tell which you are picking.

Envious Labs is not in the middle of any of these requests. Everything goes straight from your Mac to the provider, so Envious Labs never sees it either way.