Help & FAQ

Everything you need to know to install, configure, and troubleshoot the current Windows beta.

Current beta

v0.13.10 - Windows beta build

Pro activation now unlocks unlimited private dictation immediately, with more reliable long dictations and overlay navigation.

View changelog
Immediate Pro unlock - no restart after activating above the Starter limit
Reliable long dictation - complete audio with fewer truncated endings
Clear access levels - paid Pro stays separate from approved-tester Beta features

System Requirements

Minimum

  • OS: Windows 10 (64-bit) or later
  • RAM: 4 GB
  • Storage: 500 MB (+ model files)
  • CPU: Any modern x64 processor

Recommended

  • OS: Windows 11
  • RAM: 8 GB+
  • GPU: NVIDIA with CUDA (optional acceleration)
  • CPU: 4+ cores for large models

The GPU is optional. NVIDIA cards (GTX 10-series or newer) can run the GPU engine automatically. AMD and Intel graphics run on the processor for now.

Quick Start

1

Download Starter & install

Download Starter for Windows - no account or invite is required. Run the installer, then follow the Windows SmartScreen steps if a warning appears: click "More info" → "Run anyway". This is expected for new software without an EV certificate.

2

Pick a model

On first launch, PrivateTranscribe recommends a Whisper model based on your hardware. Start with the recommended one - you can always switch later in Settings → Models.

3

Set your hotkey

In the current Windows beta, the default hotkey is Ctrl+Space. You can change it in Settings → General - standard shortcuts plus Windows-native keys such as Mouse4/Mouse5, modifier-only combos, and Danish/OEM keys like ½ are supported through the native listener.

4

Start dictating

Click into any text field, press your hotkey, speak, and your transcription auto-pastes. Local transcription runs on your machine by default.

Whisper Models

All models run 100% locally. Larger models are more accurate but need more RAM and processing time.

tiny
75 MBInstant
Basic accuracy - great for quick notes
base
142 MBFast
Good for everyday dictation
small
466 MBFast
Strong accuracy for most use cases
medium
1.5 GBModerate
High accuracy, handles accents well
large-v3-turbo
1.6 GBModerate
Best speed/accuracy ratio
large-v3
3.1 GBSlower
Maximum accuracy

💡 Tip: If you have an NVIDIA GPU with CUDA, try large-v3-turbo first. PrivateTranscribe auto-detects your hardware and recommends the best option.

Key Features

Global Hotkey

Works in any app - browsers, IDEs, chat apps, documents. Press your hotkey to record, press again to transcribe and auto-paste. Tap-to-talk and push-to-talk are both supported.

Custom Dictionary

Add technical terms, names, and jargon that Whisper might not know. Your dictionary entries are used as hints during transcription.

99 Languages + Translation

Speak in your language and get text in your language. Or enable "Translate to English" to speak in any supported language and get English output.

Transcription History

Every transcription is saved locally with timestamps. Search, copy, or export your history. Track your usage streaks on the dashboard.

Frequently Asked Questions

Troubleshooting

No window appears after launch

Check the system tray (click the ^ caret near the clock). PrivateTranscribe minimizes to tray by default.


No transcription output

Make sure your microphone is selected in Settings → General and that Windows microphone permissions are enabled (Settings → Privacy → Microphone).


Hotkey doesn't work in some apps

Some apps intercept global hotkeys. Try changing your hotkey in Settings → General to avoid conflicts. On Windows, recent beta builds also support native-only hotkeys such as Mouse4/Mouse5, modifier-only combinations, and Danish/OEM keys like ½.


Need more help?

Email support@privatetranscribe.com with your system info (Settings → Developer → Copy System Info) and we'll help you out.

Ready to start?

Download Starter for Windows

5,000 words per day. No account required. Local by default, cloud optional.