Real-time meetings - API call procedure

​

Introduction

The real-time transcription APIs allow you to perform the following operations:

1. Create a real-time meeting, stream audio using the WebSocket API, and receive real-time speech recognition and translation results. This lets you implement features such as real-time captions.
2. Pause and resume meetings.
3. After a meeting ends, you can obtain an intelligent summary of the meeting.

​

Call flow

​

Create a meeting

First, call the CreateTask API to create a real-time meeting and obtain the meeting's dataId. For more information about the call procedure and configuration parameters, see CreateTask.

​

Use the WebSocket API for stream recognition

Then, stream audio using the WebSocket API to receive real-time recognition and translation results. For more information about the call procedure and configuration parameters, see Audio Transcription Interaction Protocol (WebSocket). During the meeting, you can disconnect the WebSocket connection to pause the meeting. Within the 24-hour validity period of the dataId, you can re-establish the WebSocket connection to resume the specified meeting.

​

End the meeting and create a meeting summary

After the meeting ends, call the CreateTask API to create a meeting summary analysis task. For more information about the call procedure and configuration parameters, see CreateTask. Note: You must call the CreateTask API to create a meeting summary analysis task after the meeting ends. The Tingwu Agent does not automatically create this task.

​

Get the summary generation status and results

Then, call the GetTask API to retrieve the summary generation status and results. For more information about the call procedure and return values, see GetTask.