Skip to main content
Find your problem below by its symptom. Superwhisper saves every dictation as you speak, so your recording is safe in History and can be reprocessed after the fix.
Turn on Error logging in Settings → Configuration before reproducing a problem. See Logs & reporting.

Error messages

"No voice found" or stuck processing

The recording had no detectable speech, or processing never finishes.

"LLM processing error"

Transcription worked; the AI formatting step failed. Includes API-key fixes.

App won't launch or install

Windows DLL and Vulkan errors, installer problems, macOS launch failures.

Crashes during or after transcription

Isolate whether it’s the model, the mode, or the machine.

Wrong output

Transcribed in English instead of my language

The vocabulary hint that fixes automatic language detection.

Gibberish, repeated, or made-up text

Where hallucinations come from and how to stop them.

AI answers instead of transcribing

The mode’s prompt causes this. These tips fix it.

Text not pasting, pasting twice, or pasting old text

Delivery problems: permissions, clipboard restore, simulate keypresses.

Input and controls

Shortcuts not working or misfiring

Key conflicts, global vs. per-mode shortcuts, and push-to-talk behavior.

Wrong microphone or system audio

Input switching, AirPods quality, empty meeting recordings.

Modes not saving or disappearing

Usually caused by an iCloud folder move.

Slow or heavy on the machine

Model choice, keep-loaded duration, and processing-time diagnosis.

Realtime transcription unavailable

Realtime needs a Nova cloud model or Parakeet Realtime.

Context awareness issues

How selected-text and clipboard context work, and when they backfire.

Adapting to iOS 26.4

The keyboard return-to-app change and what still works.

Logs, uninstall, and reporting

Enable logging, find your files, uninstall cleanly, report bugs well.
If your problem isn’t listed, email support@superwhisper.com with your platform, app version, and what you expected to happen.