> ## Documentation Index > Fetch the complete documentation index at: https://docs.bigmodel.cn/llms.txt > Use this file to discover all available pages before exploring further. # 音视频通话 > [GLM-Realtime](/cn/guide/models/sound-and-video/glm-realtime) 提供实时音视频通话和多模态交互能力，支持实时语音对话、视频理解、函数调用等功能。
由于浏览器安全考虑禁止 `WebSocket` 添加鉴权认证请求头，无法在此直接体验，使用详情请参考 [Realtime 指南使用](/cn/guide/models/sound-and-video/glm-realtime)。 ## AsyncAPI ````yaml asyncapi/asyncapi.json realtime id: realtime title: 音视频通话与多模态交互 description: >- [GLM-Realtime](/cn/guide/models/sound-and-video/glm-realtime) 提供实时音视频通话和多模态交互能力，支持实时语音对话、视频理解、函数调用等功能。由于浏览器安全考虑禁止 `WebSocket` 添加鉴权认证请求头，无法在此直接体验，使用详情请参考[Realtime 指南使用](/cn/guide/models/sound-and-video/glm-realtime)。 servers: - id: production protocol: wss host: open.bigmodel.cn bindings: [] variables: [] address: /api/paas/v4/realtime parameters: [] bindings: - protocol: ws version: 0.1.0 value: headers: type: object properties: Authorization: type: string description: 鉴权信息，支持JWT（客户端）或Bearer API Key（服务端）。格式：'Bearer YOUR_API_KEY' pattern: ^Bearer .+$ required: - Authorization schemaProperties: - name: headers type: object required: false properties: - name: Authorization type: string description: 鉴权信息，支持JWT（客户端）或Bearer API Key（服务端）。格式：'Bearer YOUR_API_KEY' required: true operations: - &ref_16 id: sendSessionUpdate title: Send session update description: 发送会话配置更新 type: send messages: - &ref_28 id: SessionUpdate payload: - title: 会话配置更新事件 description: 客户端发送会话配置更新 allOf: &ref_0 - &ref_1 type: object properties: event_id: type: string description: 由客户端生成的事件ID，用于标识此事件 x-parser-schema-id: type: type: string description: 事件类型 x-parser-schema-id: client_timestamp: type: integer description: 调用端发起调用的时间戳，毫秒 x-parser-schema-id: required: - type x-parser-schema-id: BaseEvent - type: object properties: type: type: string const: session.update x-parser-schema-id: session: &ref_7 title: 会话配置 description: GLM-Realtime API 的会话配置参数，用于设置模型、音频格式、语音类型、指令等参数 type: object properties: model: type: string enum: - glm-realtime-flash - glm-realtime-air description: 模型编码 x-parser-schema-id: input_audio_format: type: string description: 音频输入格式，支持wav、pcm；输入PCM格式需要增加采样率，例如pcm16、pcm24 default: wav x-parser-schema-id: output_audio_format: type: string const: pcm description: 音频输出格式，当前仅支持pcm，采样率24kHz，单声道，16位深 x-parser-schema-id: instructions: type: string description: 系统指令，用于引导模型生成期望的响应 x-parser-schema-id: voice: type: string enum: - tongtong - female-tianmei - male-qn-daxuesheng - male-qn-jingying - lovely_girl - female-shaonv default: tongtong description: 声音类型 x-parser-schema-id: temperature: type: number minimum: 0 maximum: 1 description: 模型温度，控制输出的随机性和创造性 x-parser-schema-id: max_response_output_tokens: type: string description: 回复的最大长度，对应文本token计数，范围0-1024 x-parser-schema-id: turn_detection: type: object properties: type: type: string enum: - server_vad - client_vad default: client_vad description: VAD检测类型 x-parser-schema-id: required: - type x-parser-schema-id: TurnDetection tools: type: array items: type: object properties: type: type: string const: function description: 工具类型 x-parser-schema-id: name: type: string description: 函数名称 x-parser-schema-id: description: type: string description: 函数功能描述 x-parser-schema-id: parameters: type: object description: 函数参数的JSON Schema对象 x-parser-schema-id: required: - type - name - description x-parser-schema-id: Tool description: 可用的工具列表，当前仅audio模式支持tools调用 x-parser-schema-id: modalities: type: array items: type: string enum: - text - audio x-parser-schema-id: default: - text - audio description: 输出模态，可选择文本和/或音频输出 x-parser-schema-id: input_audio_noise_reduction: type: object properties: type: type: string enum: - near_field - far_field description: 降噪类型，near_field适用于近距离麦克风，far_field适用于远距离麦克风 x-parser-schema-id: x-parser-schema-id: NoiseReduction beta_fields: type: object properties: chat_mode: type: string enum: - audio - video_passive default: audio description: 通话模式 x-parser-schema-id: tts_source: type: string const: e2e description: 文本转语音方式 x-parser-schema-id: auto_search: type: boolean default: true description: 是否开启内置的自动搜索，仅在audio模式下生效 x-parser-schema-id: required: - chat_mode x-parser-schema-id: BetaFields greeting_config: type: object properties: enable: type: boolean description: 是否开启开场白 x-parser-schema-id: content: type: string description: 开场白内容 x-parser-schema-id: x-parser-schema-id: GreetingConfig required: - input_audio_format - output_audio_format - turn_detection x-parser-schema-id: SessionConfig required: - session x-parser-schema-id: x-parser-schema-id: SessionUpdate name: 会话配置更新 headers: [] jsonPayloadSchema: title: 会话配置更新事件 description: 客户端发送此事件以配置或更新会话参数，包括模型选择、音频格式、语音类型、指令等 allOf: *ref_0 x-parser-schema-id: SessionUpdate title: 会话配置更新 description: 客户端发送会话配置更新 example: '{}' bindings: [] extensions: - id: x-parser-unique-object-id value: SessionUpdate - id: x-parser-message-name value: SessionUpdate bindings: [] extensions: &ref_3 - id: x-parser-unique-object-id value: realtime - &ref_17 id: sendAudioBuffer title: Send audio buffer description: 发送音频数据 type: send messages: - &ref_29 id: InputAudioBufferAppend payload: - title: 音频数据上传事件 description: 客户端发送音频数据 allOf: &ref_2 - *ref_1 - type: object properties: type: type: string const: input_audio_buffer.append x-parser-schema-id: audio: type: string description: 音频(wav or pcm)二进制的base64编码字符串 x-parser-schema-id: required: - audio x-parser-schema-id: x-parser-schema-id: InputAudioBufferAppend name: 音频数据上传 headers: [] jsonPayloadSchema: title: 音频数据上传事件 description: 客户端发送音频数据块到服务器，用于实时语音输入。音频数据以Base64编码格式传输 allOf: *ref_2 x-parser-schema-id: InputAudioBufferAppend title: 音频数据上传 description: 客户端发送音频数据 example: '{}' bindings: [] extensions: - id: x-parser-unique-object-id value: InputAudioBufferAppend - id: x-parser-message-name value: InputAudioBufferAppend bindings: [] extensions: *ref_3 - &ref_18 id: commitAudioBuffer title: Commit audio buffer description: 提交音频缓冲区 type: send messages: - &ref_30 id: InputAudioBufferCommit payload: - allOf: &ref_4 - *ref_1 - type: object properties: type: type: string const: input_audio_buffer.commit x-parser-schema-id: x-parser-schema-id: x-parser-schema-id: InputAudioBufferCommit name: 音频缓冲区提交 description: 客户端提交音频缓冲区 headers: [] jsonPayloadSchema: allOf: *ref_4 x-parser-schema-id: InputAudioBufferCommit title: 音频缓冲区提交 description: 客户端提交音频缓冲区 example: '{}' bindings: [] extensions: - id: x-parser-unique-object-id value: InputAudioBufferCommit - id: x-parser-message-name value: InputAudioBufferCommit bindings: [] extensions: *ref_3 - &ref_19 id: createResponse title: Create response description: 创建响应请求 type: send messages: - &ref_31 id: ResponseCreate payload: - allOf: &ref_5 - *ref_1 - type: object properties: type: type: string const: response.create x-parser-schema-id: x-parser-schema-id: x-parser-schema-id: ResponseCreate name: 创建响应请求 description: 客户端请求创建响应 headers: [] jsonPayloadSchema: allOf: *ref_5 x-parser-schema-id: ResponseCreate title: 创建响应请求 description: 客户端请求创建响应 example: '{}' bindings: [] extensions: - id: x-parser-unique-object-id value: ResponseCreate - id: x-parser-message-name value: ResponseCreate bindings: [] extensions: *ref_3 - &ref_20 id: cancelResponse title: Cancel response description: 取消响应请求 type: send messages: - &ref_32 id: ResponseCancel payload: - allOf: &ref_6 - *ref_1 - type: object properties: type: type: string const: response.cancel x-parser-schema-id: x-parser-schema-id: x-parser-schema-id: ResponseCancel name: 取消响应请求 description: 客户端取消响应 headers: [] jsonPayloadSchema: allOf: *ref_6 x-parser-schema-id: ResponseCancel title: 取消响应请求 description: 客户端取消响应 example: '{}' bindings: [] extensions: - id: x-parser-unique-object-id value: ResponseCancel - id: x-parser-message-name value: ResponseCancel bindings: [] extensions: *ref_3 - &ref_9 id: receiveSessionUpdated title: Receive session updated description: 接收会话更新确认 type: receive messages: - &ref_21 id: SessionUpdated payload: - name: 会话更新确认 description: 服务器确认会话更新 type: object properties: - name: event_id type: string required: false - name: type type: string description: session.updated required: true - name: session type: object title: 会话配置 description: GLM-Realtime API 的会话配置参数，用于设置模型、音频格式、语音类型、指令等参数 required: true properties: - name: model type: string description: 模型编码 enumValues: - glm-realtime-flash - glm-realtime-air required: false - name: input_audio_format type: string description: 音频输入格式，支持wav、pcm；输入PCM格式需要增加采样率，例如pcm16、pcm24 required: true - name: output_audio_format type: string description: 音频输出格式，当前仅支持pcm，采样率24kHz，单声道，16位深 required: true - name: instructions type: string description: 系统指令，用于引导模型生成期望的响应 required: false - name: voice type: string description: 声音类型 enumValues: - tongtong - female-tianmei - male-qn-daxuesheng - male-qn-jingying - lovely_girl - female-shaonv required: false - name: temperature type: number description: 模型温度，控制输出的随机性和创造性 required: false - name: max_response_output_tokens type: string description: 回复的最大长度，对应文本token计数，范围0-1024 required: false - name: turn_detection type: object required: true properties: - name: type type: string description: VAD检测类型 enumValues: - server_vad - client_vad required: true - name: tools type: array description: 可用的工具列表，当前仅audio模式支持tools调用 required: false properties: - name: type type: string description: 工具类型 required: true - name: name type: string description: 函数名称 required: true - name: description type: string description: 函数功能描述 required: true - name: parameters type: object description: 函数参数的JSON Schema对象 required: false - name: modalities type: array description: 输出模态，可选择文本和/或音频输出 required: false properties: - name: item type: string enumValues: - text - audio required: false - name: input_audio_noise_reduction type: object required: false properties: - name: type type: string description: 降噪类型，near_field适用于近距离麦克风，far_field适用于远距离麦克风 enumValues: - near_field - far_field required: false - name: beta_fields type: object required: false properties: - name: chat_mode type: string description: 通话模式 enumValues: - audio - video_passive required: true - name: tts_source type: string description: 文本转语音方式 required: false - name: auto_search type: boolean description: 是否开启内置的自动搜索，仅在audio模式下生效 required: false - name: greeting_config type: object required: false properties: - name: enable type: boolean description: 是否开启开场白 required: false - name: content type: string description: 开场白内容 required: false headers: [] jsonPayloadSchema: type: object properties: event_id: type: string x-parser-schema-id: type: type: string const: session.updated x-parser-schema-id: session: *ref_7 required: - type - session x-parser-schema-id: SessionUpdated title: 会话更新确认 description: 服务器确认会话更新 example: No examples found bindings: [] extensions: - id: x-parser-unique-object-id value: SessionUpdated - id: x-parser-message-name value: SessionUpdated bindings: [] extensions: *ref_3 - &ref_10 id: receiveResponseCreated title: Receive response created description: 接收响应创建确认 type: receive messages: - &ref_22 id: ResponseCreated payload: - name: 响应创建确认 description: 服务器确认响应创建 type: object properties: - name: event_id type: string required: false - name: type type: string description: response.created required: true - name: response type: object required: true properties: - name: id type: string description: 响应的唯一ID required: true - name: object type: string description: realtime.response required: true - name: status type: string description: 响应的状态 enumValues: - completed - cancelled - failed - incomplete - in_progress required: true - name: usage type: object required: false properties: - name: total_tokens type: integer description: 总共使用的令牌数量 required: false - name: input_tokens type: integer description: 输入令牌数量 required: false - name: output_tokens type: integer description: 输出令牌数量 required: false - name: input_token_details type: object required: false properties: - name: cached_tokens type: integer description: 使用缓存令牌的数量 required: false - name: text_tokens type: integer description: 使用文本令牌的数量 required: false - name: audio_tokens type: integer description: 使用音频令牌的数量 required: false - name: output_token_details type: object required: false properties: - name: text_tokens type: integer description: 输出的文本令牌数量 required: false - name: audio_tokens type: integer description: 输出的音频令牌数量 required: false headers: [] jsonPayloadSchema: type: object properties: event_id: type: string x-parser-schema-id: type: type: string const: response.created x-parser-schema-id: response: &ref_8 type: object properties: id: type: string description: 响应的唯一ID x-parser-schema-id: object: type: string const: realtime.response x-parser-schema-id: status: type: string enum: - completed - cancelled - failed - incomplete - in_progress description: 响应的状态 x-parser-schema-id: usage: type: object properties: total_tokens: type: integer description: 总共使用的令牌数量 x-parser-schema-id: input_tokens: type: integer description: 输入令牌数量 x-parser-schema-id: output_tokens: type: integer description: 输出令牌数量 x-parser-schema-id: input_token_details: type: object properties: cached_tokens: type: integer description: 使用缓存令牌的数量 x-parser-schema-id: text_tokens: type: integer description: 使用文本令牌的数量 x-parser-schema-id: audio_tokens: type: integer description: 使用音频令牌的数量 x-parser-schema-id: x-parser-schema-id: output_token_details: type: object properties: text_tokens: type: integer description: 输出的文本令牌数量 x-parser-schema-id: audio_tokens: type: integer description: 输出的音频令牌数量 x-parser-schema-id: x-parser-schema-id: x-parser-schema-id: required: - id - object - status x-parser-schema-id: RealtimeResponse required: - type - response x-parser-schema-id: ResponseCreated title: 响应创建确认 description: 服务器确认响应创建 example: No examples found bindings: [] extensions: - id: x-parser-unique-object-id value: ResponseCreated - id: x-parser-message-name value: ResponseCreated bindings: [] extensions: *ref_3 - &ref_11 id: receiveAudioDelta title: Receive audio delta description: 接收音频增量数据 type: receive messages: - &ref_23 id: ResponseAudioDelta payload: - name: 音频增量数据 description: 服务器发送音频数据流 type: object properties: - name: event_id type: string required: false - name: type type: string description: response.audio.delta required: true - name: response_id type: string description: 响应的唯一ID required: true - name: item_id type: string description: 响应会话项的ID required: true - name: output_index type: integer description: 响应中的输出项的索引 required: false - name: content_index type: integer description: 项内容数组中的内容部分的索引 required: false - name: delta type: string description: Base64编码的PCM音频数据增量，24kHz，单声道 required: true headers: [] jsonPayloadSchema: title: 音频响应增量 description: 服务器发送的音频响应增量数据，包含Base64编码的PCM音频数据块 type: object properties: event_id: type: string x-parser-schema-id: type: type: string const: response.audio.delta x-parser-schema-id: response_id: type: string description: 响应的唯一ID x-parser-schema-id: item_id: type: string description: 响应会话项的ID x-parser-schema-id: output_index: type: integer description: 响应中的输出项的索引 x-parser-schema-id: content_index: type: integer description: 项内容数组中的内容部分的索引 x-parser-schema-id: delta: type: string description: Base64编码的PCM音频数据增量，24kHz，单声道 x-parser-schema-id: required: - type - response_id - item_id - delta x-parser-schema-id: ResponseAudioDelta title: 音频增量数据 description: 服务器发送音频数据流 example: |- { "event_id": "", "type": "", "response_id": "", "item_id": "", "output_index": 123, "content_index": 123, "delta": "" } bindings: [] extensions: - id: x-parser-unique-object-id value: ResponseAudioDelta - id: x-parser-message-name value: ResponseAudioDelta bindings: [] extensions: *ref_3 - &ref_12 id: receiveTextDelta title: Receive text delta description: 接收文本增量数据 type: receive messages: - &ref_24 id: ResponseTextDelta payload: - name: 文本增量数据 description: 服务器发送文本数据流 type: object properties: - name: event_id type: string required: false - name: type type: string description: response.text.delta required: true - name: response_id type: string required: true - name: item_id type: string required: true - name: output_index type: integer required: false - name: content_index type: integer required: false - name: delta type: string description: 模型生成的文本增量 required: true headers: [] jsonPayloadSchema: title: 文本响应增量 description: 服务器发送的文本响应增量数据，包含模型生成的文本片段 type: object properties: event_id: type: string x-parser-schema-id: type: type: string const: response.text.delta x-parser-schema-id: response_id: type: string x-parser-schema-id: item_id: type: string x-parser-schema-id: output_index: type: integer x-parser-schema-id: content_index: type: integer x-parser-schema-id: delta: type: string description: 模型生成的文本增量 x-parser-schema-id: required: - type - response_id - item_id - delta x-parser-schema-id: ResponseTextDelta title: 文本增量数据 description: 服务器发送文本数据流 example: |- { "event_id": "", "type": "", "response_id": "", "item_id": "", "output_index": 123, "content_index": 123, "delta": "" } bindings: [] extensions: - id: x-parser-unique-object-id value: ResponseTextDelta - id: x-parser-message-name value: ResponseTextDelta bindings: [] extensions: *ref_3 - &ref_13 id: receiveResponseDone title: Receive response done description: 接收响应完成通知 type: receive messages: - &ref_25 id: ResponseDone payload: - name: 响应完成通知 description: 服务器发送响应完成 type: object properties: - name: event_id type: string required: false - name: type type: string description: response.done required: true - name: response type: object required: true properties: - name: id type: string description: 响应的唯一ID required: true - name: object type: string description: realtime.response required: true - name: status type: string description: 响应的状态 enumValues: - completed - cancelled - failed - incomplete - in_progress required: true - name: usage type: object required: false properties: - name: total_tokens type: integer description: 总共使用的令牌数量 required: false - name: input_tokens type: integer description: 输入令牌数量 required: false - name: output_tokens type: integer description: 输出令牌数量 required: false - name: input_token_details type: object required: false properties: - name: cached_tokens type: integer description: 使用缓存令牌的数量 required: false - name: text_tokens type: integer description: 使用文本令牌的数量 required: false - name: audio_tokens type: integer description: 使用音频令牌的数量 required: false - name: output_token_details type: object required: false properties: - name: text_tokens type: integer description: 输出的文本令牌数量 required: false - name: audio_tokens type: integer description: 输出的音频令牌数量 required: false headers: [] jsonPayloadSchema: type: object properties: event_id: type: string x-parser-schema-id: type: type: string const: response.done x-parser-schema-id: response: *ref_8 required: - type - response x-parser-schema-id: ResponseDone title: 响应完成通知 description: 服务器发送响应完成 example: No examples found bindings: [] extensions: - id: x-parser-unique-object-id value: ResponseDone - id: x-parser-message-name value: ResponseDone bindings: [] extensions: *ref_3 - &ref_14 id: receiveError title: Receive error description: 接收错误信息 type: receive messages: - &ref_26 id: ServerError payload: - name: 服务器错误 description: 服务器发送错误信息 type: object properties: - name: event_id type: string description: 服务器事件的唯一ID required: false - name: type type: string description: error required: true - name: error type: object required: true properties: - name: type type: string description: 错误类型 required: true - name: code type: string description: 错误代码 required: false - name: message type: string description: 用户可读的错误消息 required: true headers: [] jsonPayloadSchema: title: 服务器错误事件 description: 服务器发送给客户端的错误信息，包含错误类型、代码和详细消息 type: object properties: event_id: type: string description: 服务器事件的唯一ID x-parser-schema-id: type: type: string const: error x-parser-schema-id: error: type: object properties: type: type: string description: 错误类型 x-parser-schema-id: code: type: string description: 错误代码 x-parser-schema-id: message: type: string description: 用户可读的错误消息 x-parser-schema-id: required: - type - message x-parser-schema-id: required: - type - error x-parser-schema-id: ServerError title: 服务器错误 description: 服务器发送错误信息 example: |- { "event_id": "", "type": "", "error": { "type": "", "code": "", "message": "" } } bindings: [] extensions: - id: x-parser-unique-object-id value: ServerError - id: x-parser-message-name value: ServerError bindings: [] extensions: *ref_3 - &ref_15 id: receiveHeartbeat title: Receive heartbeat description: 接收心跳信号 type: receive messages: - &ref_27 id: Heartbeat payload: - name: 心跳信号 description: 服务器发送心跳 type: object properties: - name: type type: string description: heartbeat required: true headers: [] jsonPayloadSchema: title: 心跳事件 description: 服务器定期发送的心跳信号，用于保持WebSocket连接活跃状态 type: object properties: type: type: string const: heartbeat x-parser-schema-id: required: - type x-parser-schema-id: Heartbeat title: 心跳信号 description: 服务器发送心跳 example: |- { "type": "" } bindings: [] extensions: - id: x-parser-unique-object-id value: Heartbeat - id: x-parser-message-name value: Heartbeat bindings: [] extensions: *ref_3 sendOperations: - *ref_9 - *ref_10 - *ref_11 - *ref_12 - *ref_13 - *ref_14 - *ref_15 receiveOperations: - *ref_16 - *ref_17 - *ref_18 - *ref_19 - *ref_20 sendMessages: - *ref_21 - *ref_22 - *ref_23 - *ref_24 - *ref_25 - *ref_26 - *ref_27 receiveMessages: - *ref_28 - *ref_29 - *ref_30 - *ref_31 - *ref_32 extensions: - id: x-parser-unique-object-id value: realtime securitySchemes: [] ````