API

5. API Interface

The client interacts with the backend via HTTP POST requests.

5.1 Get Video

Interface Description: Negotiate and obtain WebRTC video stream
Endpoint: /offer

5.1.1 Request Parameters

Parameter	Type	Required	Description
type	string	Y	Must be `offer`
sdp	string	Y	WebRTC offer parameter
avatar	string	N	Avatar ID; default value is `avatar_id` from startup command parameters
refaudio	string	N	Cloned voice ID; default value is `REF_FILE` from startup command parameters
reftext	string	N	Text for cloned voice; default value is `REF_TEXT` from startup command parameters
custom_config	string	N	JSON string for action choreography configuration; default value is `customvideo_config` from startup command parameters
Request example:


{
    "sdp": "xxxx",
    "type": "offer"
}

5.1.2 Response Data

Parameter	Type	Required	Description
type	string	Y	Must be `answer`
sdp	string	Y	WebRTC answer parameter
sessionid	string	Y	Digital human session ID, used to distinguish multiple digital human streams. Required for all subsequent interfaces.
Response example:


{
    "sdp": "xxxx",
    "type": "answer",
    "sessionid": "d39256ea-c017-4cdc-8d4b-b73fd246d95f"
}

5.2 Send Text

Interface Description: Send chat content to the digital human
Endpoint: /human

5.2.1 Request Parameters

Parameter	Type	Required	Description
sessionid	string	Y	Digital human session ID
interrupt	bool	N	Whether to interrupt the digital human’s current speech; default is `false`
type	string	Y	`echo`: Digital human speaks the input text; `chat`: Chat with the digital human
text	string	Y	Text content
Request example:


{
    "text": "hello",
    "type": "echo",
    "interrupt": true,
    "sessionid": "d39256ea-c017-4cdc-8d4b-b73fd246d95f"
}

5.3 Switch Playback Video

Interface Description: Control the digital human to play custom video
Endpoint: /set_audiotype

5.3.1 Request Parameters

Parameter	Type	Required	Description
sessionid	string	Y	Digital human session ID
audiotype	int	Y	Video content to play, corresponding to custom videos configured on the backend
Request example:


{
    "audiotype": 2,
    "sessionid": "d39256ea-c017-4cdc-8d4b-b73fd246d95f"
}

5.4 Recording

Interface Description: Save digital human video recording on the server; file path: data/record.mp4
Endpoint: /record

5.4.1 Request Parameters

Parameter	Type	Required	Description
sessionid	string	Y	Digital human session ID
type	string	Y	`start_record`: Start recording; `end_record`: Stop recording
Request example:


{
    "type": "start_record",
    "sessionid": "d39256ea-c017-4cdc-8d4b-b73fd246d95f"
}

5.5 Interrupt Digital Human Speech

Interface Description: Interrupt the digital human’s current speech
Endpoint: /interrupt_talk

5.5.1 Request Parameters

Parameter	Type	Required	Description
sessionid	string	Y	Digital human session ID
Request example:


{
    "sessionid": "d39256ea-c017-4cdc-8d4b-b73fd246d95f"
}

Third-Party Software Integration Process


pc = new RTCPeerConnection();
var offer = pc.localDescription;
Call the API `/offer` with parameters `offer.sdp` and `offer.type`.
Obtain the response data `answer`.
pc.setRemoteDescription(answer); // Set the response to the PC object; video stream will be received afterward.
// For detailed code, refer to web/client.js

Call the API `/human` with `sessionid` from `answer.sessionid` to trigger the digital human to speak.

Docker Virtual Camera Usage