CafeTran

Note: This is a solution for continuous voice dictation, not voice editing, on GNU/Linux. It cannot compare to integrated workflows (voice dictation, voice editing, voice commands) such as Dragon NaturallySpeaking on PC/Mac or Apple dictation on Mac. The following does not address any privacy concern, which should be checked against the relevant terms of use. 

The idea is to use the microphone of an Android smartphone (which is optimized for voice input and makes voice recognition a lot easier without the need for additional special equipment ie. PC microphone) along with Dragon dictation as integrated in the Swype keyboard app (https://play.google.com/store/apps/details?id=com.nuance.swype.dtc&hl=en), since Dragon is not readily usable in Linux.

Steps:
- Install Swype keyboard (integrating Dragon Dictation) on Android smartphone
- Use Max Remote to use smartphone as a remote and establish a connection with the PC. Other remote apps may use Bluetooth. This one requires to be connected on the same Wi-Fi network with both the PC and the smartphone.

Download:
Android app: https://play.google.com/store/apps/details?id=com.bitunits.maxremote&hl=enT
Server app (portable Java version for Linux): http://www.bitunitsstudio.com/max-remote-server.html

Tutorial for server
http://www.bitunitsstudio.com/tutorial-como-usar-o-max-remote.html

After the Wi-Fi connection and the installation of Max Remote on both devices:

– Enable Swype and set it as the default keyboard on Android, add the languages needed.
– Launch Max Remote on PC to activate server
– Open Max Remote on smartphone while connected on same W-ifi network
-Add server (Gear icon>Server>Plus icon)
-Select the keypad icon
– On the bottom, select the Keyboard, NOT the Microphone (as it uses Google voice, not Dragon)
-Long press the Space key to select the input language
-Press the microphone button to launch Dragon in Swype keyboard. Make sure the language code shown is the correct one.

The microphone is on until you stop talking. What you say is typed directly on the text aware application you are focusing on the GNU/Linux PC. 

Advice: To have the microphone open for continuous recognition: disable Detect end-of-speech in Swype settings (Language & Input > Swype&Dragon >Settings). Thusly, the voice recognition is always on, in the selected language. Whatever you say will be directly be typed in the application.

Works fine in CafeTran. Could also be used with "Bind external editor" or simply work independently, directly voice typing in the CAT tool, or across any application.

IK: The Google Docs Voice Typing tool in Google Chrome browser is more straightforward.

There are people who try to avoid using products of that advertising company. Me, for instance.

H.