Pick a model group to try out Speech to Text and Text to
Speech.
Speak naturally. Watch the transcript arrive with
live timestamps.
Connect to a live session and watch partial and
final recognition arrive as audio is processed.
Idle
Session not connected
00:00
Transcript
Recognition will appear here once the session
starts.
Timestamps
Live word timestamps will appear here.
Microphone access is required for the live demo.
Type text and hear it spoken in your language.
Select a language, enter the text you want to
synthesise, then click
Generate Speech to hear the result.
 scale(0.125)%22%3E%3Cpath fill=%22%23000%22 d=%22M640,0C286.5,0,0,143.3,0,320s286.5,320,640,320s640-143.3,640-320S993.5,0,640,0z M640,589.5C314.4,589.5,50.5,468.8,50.5,320S314.4,50.5,640,50.5s589.5,120.7,589.5,269.5S965.6,589.5,640,589.5z M559.3,263.9v-50.5h180.5v50.5h-65v162.8h-50.5V263.9H559.3z M957.8,213.3v213.3h-50.5l-91.6-127v127h-50.5V213.3h50.5l91.6,127v-127L957.8,213.3z M320.7,426.7V213.3h50.5l56.1,86.3l56.1-86.3H534v213.3h-50.5V306l-38.3,58.9h-35.6L371.2,306v120.7H320.7z%22/%3E%3C/g%3E%3C/svg%3E)
