After you configure the application, test its real-time performance. Once you confirm that it works correctly, you can publish the application and call it from your online service.
Application experience
After you complete the configuration, click Run Now to test the application.
For the best conversation experience, wear headphones and test in a quiet environment.
-
To configure an API key, follow these steps.
-
If authorization has not been granted, grant it. After the authorization is successful, create an API key.
-
If authorization has already been granted, you can create an API key directly.
-
-
Enable interactive voice response to test real-time voice conversations. Disable it to test text conversations. In a multimodal application, you can also select a real-time video chat agent to test real-time video conversations.
-
Before you test the application for the first time, you must grant the page permission to use the camera and microphone.
-
If you did not enable interactive voice response during the application configuration, only text interaction is supported in the test panel on the right.
-
The console displays the end-to-end latency. First-packet end-to-end latency is the time from when Automatic Speech Recognition (ASR) detects the last word to when Text-to-Speech (TTS) plays the first word. This latency is the wait time that the user experiences.
-
Publish the application
After you confirm the application's performance, click Publish. In the Published Version dialog box, enter a Version Description and click OK. The current configuration is then published. You can then use the application's unique ID to develop your services.
After you publish the application, you can continue to modify its configuration. These changes do not affect the online service until you publish a new version.