superwhisper pasted raw text: the AI step failed, and here is how to get the formatted version

When superwhisper's AI step fails on a Mac, it pastes your raw transcription in place of the mode's formatted output. Nothing is lost: the dictation is saved in History, and you can reprocess it once the cause is fixed. The cause is usually the connection, because most language models run in the cloud even when the voice model is local. On Windows the same failure shows as an error message.

I built YapToText, a dictation app for the Mac, so read what I say about superwhisper with that in mind. Every fact about superwhisper here comes from its own documentation, read on October 11, 2026.

Download YapToText on the Mac App StoreThis app is completely free. You don't have to make an account. There are no subscriptions, no difficult downloads, and nothing is locked. Made for Apple silicon Macs. Source on GitHub.

Fix it in four steps

  1. Check your connection. If you are offline, reconnect, or switch the mode's Language Model to a local one. The docs name S1-mini.
  2. Open the dictation in History and reprocess it. A passing provider error clears on a retry.
  3. Edit the mode and pick a different Language Model. If the error follows one model, that provider is the problem.
  4. Update superwhisper with Check for Updates in Settings, Configuration.

A mode has two models: a voice model that hears you and a language model that formats. A local voice model does not make the language model local. That is why a mode can work offline for transcription and still fail to format.

The Windows messages

On Windows you may see Error during local LLM inference, or Error while making chat query. superwhisper's docs say the same four fixes apply.

If the failing model uses your own API key

  • Paste the key again, fresh from the provider's dashboard. Expired or restricted keys fail.
  • Create a new key if one keeps failing, and paste that.
  • Check the provider's status. An outage there makes the AI step fail.
  • Or stop using your own key. The built-in cloud models need none of this. Switch the mode back to one.

Your own keys are under Settings, Models library, Bring your own keys.

Report it

  1. Turn on Error logging in Settings, Configuration.
  2. Reproduce the error once.
  3. Report the recording from History, and name the mode and the model it uses.

The superwhisper history page covers reprocessing, and the modes page covers the two models a mode carries.

Cleanup that cannot go offline

This failure exists wherever cleanup depends on a cloud model. YapToText, the free Mac app I built, runs its cleanup model on the Mac, inside the App Store download, so there is no connection or provider for that step to lose. If you want the exact words with no cleanup at all, Raw Transcription is one of its modes. The trade is that a small on-device model does less than the large cloud models superwhisper can reach. superwhisper offline covers what its local models do without a connection.

Download YapToText on the Mac App StoreThis app is completely free. You don't have to make an account. There are no subscriptions, no difficult downloads, and nothing is locked. Made for Apple silicon Macs. Source on GitHub.

The short version

  • What happened. The AI step failed, so the raw transcription was pasted.
  • Your words. Safe in History.
  • Fix. Connection, reprocess, another language model, update.
  • Own keys. Paste a fresh key, or use a built-in model.
  • Offline. Pick a local language model for that mode.

If superwhisper changes these steps, open an issue on GitHub and I will update this page. Every one gets read.

Download YapToText on the Mac App StoreThis app is completely free. You don't have to make an account. There are no subscriptions, no difficult downloads, and nothing is locked. Made for Apple silicon Macs. Source on GitHub.

About the author. Ryleigh Newman is a developer whose hands make typing difficult and who dictates every day with YapToText, the app this site describes. Facts about other apps come from their own sites, with the date they were read. More about the project.

Sources

Questions

Why did superwhisper paste unformatted text?

The AI processing step failed, so it pasted your raw transcription. The dictation is saved in History and can be reprocessed.

How do I reprocess a dictation in superwhisper?

Open the dictation in History and reprocess it, after fixing the connection or switching the mode's language model.

Why does superwhisper AI fail when I am offline?

Most language models run in the cloud, even in a mode with a local voice model. Switch the mode's Language Model to a local one.

What does Error while making chat query mean in superwhisper?

It is the Windows form of the same failure: the language model step did not complete. Check the connection, retry from History, or change the model.