Mozilla extension testing with the implementation of an experimental voice navigation system, allowing the use of voice commands for performing standard actions in the browser. Currently, only English language commands are supported. To activate it, users need to click the indicator in the address bar and issue a voice command (the microphone is off by default).
Unlike typical voice control systems, the proposed extension does not focus on replacing the mouse and keyboard for interface manipulation, but is positioned as an assistive tool for processing natural language queries, serving as a voice assistant. For example, users can issue commands such as "what's the weather now," "find the tab with Gmail," "mute," "save as PDF," "zoom in," "open the Mozilla website."
After installing the extension, users are asked to provide permission for collecting and analyzing voice patterns, which are sent to Mozilla's servers to improve the accuracy of the service (data is collected anonymously and not shared with third parties). The submission of telemetry data with voice inputs is optional and can be declined.
According to According to the publication soeren-hentzschel.at, command processing utilizes Google’s speech recognition service (Google Cloud Speech Service), but within the extension's code, Mozilla's servers (settings can be overridden during compilation). In the file defining the privacy policy, the possibility of sending voice data to both Mozilla and Google Cloud Speech.
Source: opennet.ru
