1. Install and sign in
Download the Windows x64 installer from the official download page, open the app and sign in. You need internet access and available usage allowance. The Free plan lets you begin without buying a paid subscription.
Whispek 0.1.11 supports English and Simplified Chinese interfaces. Choose on the sign-in screen or under Settings → General → Display language. Your saved choice takes priority; Automatic follows the first supported system language and uses English when none matches. The website and app remember their language choices separately.
2. Check the microphone and language
Open Settings（设置）and Voice & input（语音与输入）. Select your microphone and make sure Windows allows microphone access. Use automatic recognition or choose the language preferences offered by the app.
For English dictation, choose English if you prefer an explicit language hint. Available languages are loaded from the service; a language selection is a recognition preference, not a guarantee of accuracy.
3. Make a short recording
Open a plain text editor and place the cursor in an editable area. Check the current voice shortcut on the Whispek home screen; the Windows default is Right Alt. Start recording, speak a short sentence and stop.
Try: “Please send the updated draft tomorrow morning.” Look for the result in the text field. Check that the meaning and punctuation are what you intended before moving on.
4. Review the text and add recurring terms
Whispek prepares the recognized speech with text cleanup. Read the result, check names and numbers, and edit it in your destination app before sending or saving. Add recurring names and specialist terms in Personal dictionary（个性词典）.
Do not try to fix every issue by adding a dictionary entry. If ordinary words are also wrong, check microphone placement, background noise and the selected input device first.
5. Know how to recover
- No recording: check microphone permission, the selected device and its input level.
- Shortcut does nothing: verify the binding and change it if another app uses it.
- Processing fails: check the connection, sign-in and remaining usage; read the app’s error message.
- Text exists but did not insert: copy the available result and paste it into the intended field.
- Need a previous result: look in History（历史记录）if that recording was retained.
6. Choose history and learning controls
Voice history is local and defaults to 50 retained records. You can change the retention setting or clear records. Assistant conversations use separate controls.
Cloud dictionary learning is enabled by default and can be turned off in dictionary settings. Read the privacy and data-control pages before dictating information that needs special handling.