Skip to main content
Superwhisper runs natively on Windows 10 and 11. It’s the same app and the same license as macOS; one purchase covers both platforms.

​
Install

  1. Download the installer from superwhisper.com/windows
  2. Run the installer and follow the prompts
  3. Superwhisper starts and adds an icon to your system tray. Click it to open settings
  4. Allow microphone access when Windows asks
  5. Follow the setup wizard, then try your first dictation (Quickstart)
Bought the Mac version by mistake, or switching machines? Download the Windows installer and enter the same key.

​
Requirements

  • Windows 10 or Windows 11 (64-bit)
  • Local voice models run on your GPU, and support depends on your hardware:
    • Cohere Transcribe and Whisper Large v3 Turbo run through Vulkan. They need a Vulkan-capable GPU with current drivers, and they’re the fastest options on machines that have one.
    • Parakeet Multilanguage runs through ONNX Runtime and works on any Windows machine, including ARM (Snapdragon) devices. It trades some speed for that compatibility.
    • On Windows ARM, Cohere Transcribe and Whisper Large v3 Turbo aren’t supported yet. Use Parakeet Multilanguage or a cloud model.
  • Cloud models run on any machine. Use them when your GPU can’t run local models, such as old integrated graphics or a virtual machine.

​
Installation problems

The app won’t launch, or you see an error like vkGetPhysicalDeviceFeatures2 could not be located or a missing DLL.
  1. Update your graphics drivers from your GPU vendor (Intel, NVIDIA, or AMD)
  2. Restart and launch Superwhisper again
  3. If the app opens but local models fail with this error, your GPU may not support Vulkan. Pick a cloud model instead; cloud models work without a GPU
  4. If you’re still stuck, email support@superwhisper.com with your Windows version and GPU model

​
Where your files live

Superwhisper stores all configuration, recordings, and modes in:
Paste that path into the File Explorer address bar to open it. You can’t change this location on Windows yet.

​
Differences from macOS

Most features work the same. These aren’t available on Windows yet:
  • Filesync: syncing configuration across installations (in development)
  • Custom app folder location: the default path above is always used
  • Hold Shift to Auto-Send
  • Simulate Keypresses
  • Coding agent integrations (Claude Code, OpenCode, Codex, Cursor, Pi, Grok, DeepSeek) and the CLI
Everything else works the same as on macOS, so the rest of these docs apply to both platforms. That includes modes and custom modes, vocabulary and replacements, local and cloud models, file transcription, realtime transcription, system audio recording, and speaker separation.

​
Next steps

Quickstart

First dictation in less than a minute

Create your first mode

Set up a Message mode and pick a model your hardware supports