Server events for Qwen-ASR-Realtime - Alibaba Cloud Model Studio

User guide: For model overview, features, and complete sample code, see Real-time speech recognition - Qwen.

error

Sent when the server detects a client or server error.

Parameter	Type	Description
type	string	The event type. Fixed to `error`.
event_id	string	The event ID.
error.type	string	The error type.
error.code	string	The error code.
error.message	string	The specific error message. For solutions, see Error codes.
error.param	string	The parameter related to the error.
error.event_id	string	The event ID related to the error.

{
  "event_id": "event_B2uoU7VOt1AAITsPRPH9n",
  "type": "error",
  "error": {
    "type": "invalid_request_error",
    "code": "invalid_value",
    "message": "Invalid value: 'whisper-1xx'. Supported values are: 'whisper-1'.",
    "param": "session.input_audio_transcription.model",
    "event_id": "event_123"
  }
}

session.created

Sent first after successful connection. Contains default session configurations.

Parameter	Type	Description
type	string	The event type. Fixed to `session.created`.
event_id	string	The event ID.
session.id	string	The ID of the current WebSocket session.
session.object	string	Fixed to `realtime.session`.
session.model	string	The model name.
session.modalities	array[string]	The output modality of the model. Fixed to `["text"]`.
session.input_audio_format	string	The input audio format.
session.input_audio_transcription	object	Configuration for speech recognition. See `input_audio_transcription` in the client's session.update event for details.
session.turn_detection	object	The Voice Activity Detection (VAD) configuration.
session.turn_detection.type	string	Fixed to `server_vad`.
session.turn_detection.threshold	float	The VAD detection threshold.
session.turn_detection.silence_duration_ms	integer	The VAD sentence-break detection threshold in milliseconds (ms).

{
    "event_id": "event_1234",
    "type": "session.created",
    "session": {
        "id": "sess_001",
        "object": "realtime.session",
        "model": "qwen3-asr-flash-realtime",
        "modalities": ["text"],
        "input_audio_format": "pcm16",
        "input_audio_transcription": null,
        "turn_detection": {
            "type": "server_vad",
            "threshold": 0.5,
            "silence_duration_ms": 200
        }
    }
}

session.updated

Sent after successfully processing the client's session.update event. On error, an error event is sent instead.

Parameter	Type	Description
type	string	The event type. The value is `session.updated`.

For descriptions of the other parameters, see session.created.

{
    "event_id": "event_1234",
    "type": "session.updated",
    "session": {
        "id": "sess_001",
        "object": "realtime.session",
        "model": "gpt-4o-realtime-preview-2024-12-17",
        "modalities": ["text"],
        "input_audio_format": "pcm16",
        "input_audio_transcription": null,
        "turn_detection": {
            "type": "server_vad",
            "threshold": 0.5,
            "silence_duration_ms": 200
        }
    }
}

input_audio_buffer.speech_started

Sent in VAD mode when speech starts in the audio buffer. Can occur each time audio is added to the buffer, unless speech start has been detected.

Parameter	Type	Description
type	string	The event type. Fixed to `input_audio_buffer.speech_started`.
event_id	string	The event ID.
audio_start_ms	integer	Milliseconds from when audio started writing to the buffer until speech was first detected in the session.
item_id	string	The ID of the user message item that will be created.

{
  "event_id": "event_B1lV7FPbgTv9qGxPI1tH4",
  "type": "input_audio_buffer.speech_started",
  "audio_start_ms": 64,
  "item_id": "item_B1lV7jWLscp4mMV8hSs8c"
}

input_audio_buffer.speech_stopped

Sent in VAD mode when speech ends in the audio buffer. After this event, the server immediately sends a conversation.item.created event containing the user message item created from the audio buffer.

Parameter	Type	Description
type	string	The event type. Fixed to `input_audio_buffer.speech_stopped`.
event_id	string	The event ID.
audio_end_ms	integer	Milliseconds elapsed from session start to when speech stopped.
item_id	string	ID of the user message item created when speech stops.

{
  "event_id": "event_B3GGEYh2orwNIdhUagZPz",
  "type": "input_audio_buffer.speech_stopped",
  "audio_end_ms": 28128,
  "item_id": "item_B3GGE8ry4yqbqJGzrVhEM"
}

input_audio_buffer.committed

VAD mode: Sent after the client sends audio data via input_audio_buffer.append event.
Manual mode: Sent after the client sends audio data via input_audio_buffer.append event, then sends input_audio_buffer.commit event.

Parameter	Type	Description
type	string	The event type. Fixed to `input_audio_buffer.committed`.
event_id	string	The event ID.
previous_item_id	string	The ID of the previous conversation item.
item_id	string	The ID of the user conversation item to be created.

{
    "event_id": "event_1121",
    "type": "input_audio_buffer.committed",
    "previous_item_id": "msg_001",
    "item_id": "msg_002"
}

conversation.item.created

Sent when a new conversation item is created.

Parameter	Type	Description
type	string	The type of the event. Fixed to `conversation.item.created`.
event_id	string	The event ID.
previous_item_id	string	The ID of the previous conversation item.
item	object	The item to add to the conversation.
item.id	string	The unique ID of the conversation item.
item.object	string	Fixed to `realtime.item`.
item.type	string	Fixed to `message`.
item.status	string	The status of the conversation item.
item.role	string	The role of the message sender.
item.content	array[object]	The content of the message.
item.content.type	string	Fixed to `input_audio`.
item.content.transcript	string	Fixed to `null`. The complete recognition result is provided in the conversation.item.input_audio_transcription.completed event.

{
  "type": "conversation.item.created",
  "event_id": "event_B3GGKbCfBZTpqFHZ0P8vg",
  "previous_item_id": "item_B3GGE8ry4yqbqJGzrVhEM",
  "item": {
    "id": "item_B3GGEPlolCqdMiVbYIf5L",
    "object": "realtime.item",
    "type": "message",
    "status": "completed",
    "role": "user",
    "content": [
      {
        "type": "input_audio",
        "transcript": null
      }
    ]
  }
}

conversation.item.input_audio_transcription.text

Sent frequently with real-time recognition results.

Parameter	Type	Description
type	string	The type of the event. Fixed to `conversation.item.input_audio_transcription.text`.
event_id	string	The ID of the event.
item_id	string	The ID of the associated conversation item.
content_index	integer	The index of the content part that contains the audio.
language	string	Language of the recognized audio. Matches the `language` request parameter if specified. Possible values: zh: Chinese (Mandarin, Sichuanese, Minnan, and Wu) yue: Cantonese en: English ja: Japanese de: German ko: Korean ru: Russian fr: French pt: Portuguese ar: Arabic it: Italian es: Spanish hi: Hindi id: Indonesian th: Thai tr: Turkish uk: Ukrainian vi: Vietnamese
emotion	string	The emotion detected in the audio. The following emotions are supported: `surprised` `neutral` `happy` `sad` `disgusted` `angry` `fearful`
text	string	Confirmed text prefix — the part of the current sentence that the model has verified and won't change.
stash	string	Pre-recognized text suffix following the confirmed part. Temporary draft the model is still processing and may correct.

{
  "event_id": "event_R7Pfu8QVBfP5HmpcbEFSd",
  "type": "conversation.item.input_audio_transcription.text",
  "item_id": "item_MpJQPNQzqVRc9aC9zMwSj",
  "content_index": 0,
  "language": "en",
  "emotion": "neutral",
  "text": "",
  "stash": "Beijing's"
}

To get the complete sentence preview, concatenate: text + stash.

Click to view an example

For example, assume a user says, "The weather is nice today, sunny and bright."The following table shows the event stream you might receive and explains how to interpret it:

Timestamp	User speech progress	API response (text and stash)	UI display (text + stash)
T1	"The..."	text: "" stash: "The"	The
T2	"...weather is..."	Text: "" stash: "The weather is"	The weather is
T3	"...nice today"	text: "The" stash: "weather is nice today"	The weather is nice today (Note: "The" is confirmed and moved to the text field.)
T4	(Short pause)	text: "The weather is nice today," stash: ""	The weather is nice today, (The first clause is fully confirmed.)
T5	"...sunny and..."	text: "The weather is nice today," stash: "sunny and"	The weather is nice today, sunny and
T6	"...bright."	text: "The weather is nice today," stash: "sunny and bright."	The weather is nice today, sunny and bright.
T7	(User stops speaking)	-	Use the content of the transcript from the conversation.item.input_audio_transcription.completed event as the final result.

conversation.item.input_audio_transcription.completed

Sends the final recognition result, marking the end of a conversation item.

Parameter	Type	Description
type	string	The event type. Fixed to `conversation.item.input_audio_transcription.completed`.
event_id	string	The event ID.
item_id	string	The ID of the associated conversation item.
content_index	integer	The index of the content part that contains the audio.
language	string	Language of the recognized audio. Matches the `language` request parameter if specified. Possible values: zh: Chinese (Mandarin, Sichuanese, Minnan, and Wu) yue: Cantonese en: English ja: Japanese de: German ko: Korean ru: Russian fr: French pt: Portuguese ar: Arabic it: Italian es: Spanish hi: Hindi id: Indonesian th: Thai tr: Turkish uk: Ukrainian vi: Vietnamese
emotion	string	The emotion detected in the audio. The following emotions are supported: `surprised` `neutral` `happy` `sad` `disgusted` `angry` `fearful`
transcript	string	The transcription result.

{
  "event_id": "event_B3GGEjPT2sLzjBM74W6kB",
  "type": "conversation.item.input_audio_transcription.completed",
  "item_id": "item_B3GGC53jGOuIFcjZkmEQ9",
  "content_index": 0,
  "language": "en",
  "emotion": "neutral",
  "transcript": "What's the weather like today?"
}

conversation.item.input_audio_transcription.failed

Sent when input audio recognition fails. Handled separately from other error events to identify the specific failed item.

Parameter	Type	Description
type	string	The event type. Fixed to `conversation.item.input_audio_transcription.failed`.
item_id	string	The ID of the associated conversation item.
content_index	integer	The index of the content part that contains the audio.
error.code	string	The error code.
error.message	string	The error message.
error.param	string	The parameter related to the error.

{
  "type": "conversation.item.input_audio_transcription.failed",
  "item_id": "<item_id>",
  "content_index": 0,
  "error": {
    "code": "<code>",
    "message": "<message>",
    "param": "<param>"
  }
}

session.finished

Session finished; all audio recognition complete. Sent after the client sends session.finish event. Client can disconnect after receiving this.

Parameter	Type	Description
type	string	The event type. The value is `session.finished`.
event_id	string	The event ID.

{
  "event_id": "event_2239",
  "type": "session.finished"
}

​error

​session.created

​session.updated

​input_audio_buffer.speech_started

​input_audio_buffer.speech_stopped

​input_audio_buffer.committed

​conversation.item.created

​conversation.item.input_audio_transcription.text

​conversation.item.input_audio_transcription.completed

​conversation.item.input_audio_transcription.failed

​session.finished