Support

First, download a Whisper model during onboarding or in Settings > Transcription. Then press Ctrl+Shift+Space (or your custom hotkey) to start recording. Press it again to stop, and your speech will be transcribed and inserted at your cursor.

Start with Small (~460 MB) for a good balance of speed and accuracy. Base (~140 MB) is faster but slightly less accurate. Medium (~1.5 GB) and Large-v3 (~3 GB) offer the best accuracy but require more memory and processing time.

After transcription, Wspr can optionally rewrite your text to fix grammar, adjust tone, or match a style. Use cloud providers like OpenAI, Anthropic, or Gemini with your own API key. Built-in rewrite styles are free; custom styles require Pro.

No. All speech-to-text transcription happens entirely on your device using whisper.cpp with GPU acceleration. Audio never leaves your computer. Cloud AI is only used for optional text rewriting (not transcription) and only if you choose to configure it.

Wspr is available as a .deb package (Ubuntu, Debian, Mint, Pop!_OS) and a .pacman package (Arch, Manjaro, EndeavourOS). Both packages include all dependencies automatically — just install and go.

This means xdotool is not installed. Wspr uses xdotool to simulate a paste keystroke in your focused application. Install it with your package manager (see above), then restart Wspr.

Open Wspr Settings and go to the License tab. Enter your license key to activate Pro. Your license key was emailed to you after purchase and can be used on any device.

Try a larger model (Small or Medium) for better accuracy. Make sure your microphone is working and selected correctly in System Settings. Speaking clearly and minimizing background noise also helps.

Wspr supports 99 languages through OpenAI's Whisper models, including English, French, Spanish, German, Japanese, Chinese, Arabic, and many more. Language is detected automatically, or you can set a preferred language in Settings.