Copy the license key from your old version. You can find this easily by saying "open setup".
Download and install the new version, then start it up.
Supply your license key to the new version when it starts up.
If you have made customizations to the old the version settings, apply those custom settings to the new version.
Delete the old version once you are comfortable that you no longer need it.
Changes:
Added support for the Whisper recognizer. There has been a lot of hype for this recognizer. While it is accurate, it is notably slower than Vosk, so we are keeping Vosk the default. To try this recognizer, you just need to set the "recognizer" field in your config/settings.yaml file to "whispercpp".
Updated the Deepgram recognizer so other models besides the default can be selected. We also changed the default model to "nova-2", which has better accuracy and performance than "enhanced-general" for dictation and voice commands. If you need medical dictation, the "nova-2-medical" model seems to perform well also.
Added new commands "utterance spacing on" and "utterance spacing off" to control whether automatic spacing is appended to each utterance. See "Auto Utterance Spacing" in the help documentation for interpreter states.
Added new commands "capitalization on" and "capitalization off" to control whether automatic capitalization provided by the recognizer is enabled. See "Auto Capitalization" in the help documentation for interpreter states. Note that the default recognizer, Vosk, does not provide capitalization, but every other supported recognizer does.
Added an "alternates" field to the numberWords setting in number settings to simplify updates.