Realtime translation client events

These are events that the OpenAI Realtime Translation WebSocket server will accept from the client.

session.update

Send this event to update the translation session configuration. Translation sessions support updates to audio.output.language, audio.input.transcription, and audio.input.noise_reduction.

Schema

Schema name: RealtimeTranslationClientEventSessionUpdate

{
  "(resource) realtime > (model) realtime_translation_session_update_event > (schema)": {
    "kind": "HttpDeclTypeAlias",
    "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionUpdate",
    "ident": "RealtimeTranslationSessionUpdateEvent",
    "type": {
      "kind": "HttpTypeObject",
      "members": [
        {
          "ident": "session"
        },
        {
          "ident": "type"
        },
        {
          "ident": "event_id"
        }
      ]
    },
    "docstring": "Send this event to update the translation session configuration. Translation\nsessions support updates to `audio.output.language`, `audio.input.transcription`,\nand `audio.input.noise_reduction`.\n",
    "childrenParentSchema": "object",
    "children": [
      "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) session",
      "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) type",
      "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) event_id"
    ]
  },
  "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) session": {
    "kind": "HttpDeclProperty",
    "title": "Realtime translation session update",
    "docstring": "Translation session fields to update. The session `type` and `model` are set\nat creation and cannot be changed with `session.update`.\n",
    "key": "session",
    "optional": false,
    "nullable": false,
    "type": {
      "kind": "HttpTypeReference",
      "ident": "RealtimeTranslationSessionUpdateRequest",
      "$ref": "(resource) realtime > (model) realtime_translation_session_update_request > (schema)"
    },
    "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionUpdate/properties/session",
    "deprecated": false,
    "schemaType": "object",
    "modelImplicit": false,
    "modelPath": "(resource) realtime > (model) realtime_translation_session_update_request",
    "childrenParentSchema": "object",
    "children": [
      "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio"
    ]
  },
  "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) type": {
    "kind": "HttpDeclProperty",
    "docstring": "The event type, must be `session.update`.",
    "key": "type",
    "optional": false,
    "nullable": false,
    "type": {
      "kind": "HttpTypeUnion",
      "types": [
        {
          "kind": "HttpTypeLiteral",
          "literal": "session.update"
        }
      ],
      "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionUpdate/properties/type"
    },
    "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionUpdate/properties/type",
    "deprecated": false,
    "schemaType": "enum",
    "childrenParentSchema": "enum",
    "children": [
      "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) type > (member) 0"
    ]
  },
  "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) event_id": {
    "kind": "HttpDeclProperty",
    "docstring": "Optional client-generated ID used to identify this event.",
    "key": "event_id",
    "optional": true,
    "nullable": false,
    "type": {
      "kind": "HttpTypeString"
    },
    "constraints": {
      "maxLength": 512
    },
    "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionUpdate/properties/event_id",
    "deprecated": false,
    "schemaType": "string",
    "children": []
  },
  "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio": {
    "kind": "HttpDeclProperty",
    "docstring": "Configuration for translation input and output audio.\n",
    "key": "audio",
    "optional": true,
    "nullable": false,
    "type": {
      "kind": "HttpTypeObject",
      "members": [
        {
          "ident": "input"
        },
        {
          "ident": "output"
        }
      ]
    },
    "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio",
    "deprecated": false,
    "schemaType": "object",
    "childrenParentSchema": "object",
    "children": [
      "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input",
      "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) output"
    ]
  },
  "(resource) realtime > (model) realtime_translation_session_update_request > (schema)": {
    "kind": "HttpDeclTypeAlias",
    "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest",
    "ident": "RealtimeTranslationSessionUpdateRequest",
    "type": {
      "kind": "HttpTypeObject",
      "members": [
        {
          "ident": "audio"
        }
      ]
    },
    "docstring": "Realtime translation session fields that can be updated with `session.update`.\n",
    "childrenParentSchema": "object",
    "children": [
      "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio"
    ]
  },
  "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) type > (member) 0": {
    "kind": "HttpDeclReference",
    "type": {
      "kind": "HttpTypeLiteral",
      "literal": "session.update"
    }
  },
  "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input": {
    "kind": "HttpDeclProperty",
    "key": "input",
    "optional": true,
    "nullable": false,
    "type": {
      "kind": "HttpTypeObject",
      "members": [
        {
          "ident": "noise_reduction"
        },
        {
          "ident": "transcription"
        }
      ]
    },
    "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio/properties/input",
    "deprecated": false,
    "schemaType": "object",
    "childrenParentSchema": "object",
    "children": [
      "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) noise_reduction",
      "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) transcription"
    ]
  },
  "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) output": {
    "kind": "HttpDeclProperty",
    "key": "output",
    "optional": true,
    "nullable": false,
    "type": {
      "kind": "HttpTypeObject",
      "members": [
        {
          "ident": "language"
        }
      ]
    },
    "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio/properties/output",
    "deprecated": false,
    "schemaType": "object",
    "childrenParentSchema": "object",
    "children": [
      "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) output > (property) language"
    ]
  },
  "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) noise_reduction": {
    "kind": "HttpDeclProperty",
    "docstring": "Optional input noise reduction. Set to `null` to disable it.\n",
    "key": "noise_reduction",
    "optional": true,
    "nullable": true,
    "type": {
      "kind": "HttpTypeObject",
      "members": [
        {
          "ident": "type"
        }
      ]
    },
    "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio/properties/input/properties/noise_reduction",
    "deprecated": false,
    "schemaType": "object",
    "childrenParentSchema": "object",
    "children": [
      "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) noise_reduction > (property) type"
    ]
  },
  "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) transcription": {
    "kind": "HttpDeclProperty",
    "docstring": "Optional source-language transcription. When configured, the server emits\n`session.input_transcript.delta` events. Translation itself still runs from\nthe input audio stream.\n",
    "key": "transcription",
    "optional": true,
    "nullable": true,
    "type": {
      "kind": "HttpTypeObject",
      "members": [
        {
          "ident": "model"
        }
      ]
    },
    "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio/properties/input/properties/transcription",
    "deprecated": false,
    "schemaType": "object",
    "childrenParentSchema": "object",
    "children": [
      "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) transcription > (property) model"
    ]
  },
  "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) output > (property) language": {
    "kind": "HttpDeclProperty",
    "docstring": "Target language for translated output audio and transcript deltas.\n",
    "key": "language",
    "optional": true,
    "nullable": false,
    "type": {
      "kind": "HttpTypeString"
    },
    "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio/properties/output/properties/language",
    "deprecated": false,
    "schemaType": "string",
    "children": []
  },
  "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) noise_reduction > (property) type": {
    "kind": "HttpDeclProperty",
    "docstring": "Type of noise reduction. `near_field` is for close-talking microphones such as headphones, `far_field` is for far-field microphones such as laptop or conference room microphones.\n",
    "key": "type",
    "optional": false,
    "nullable": false,
    "type": {
      "kind": "HttpTypeReference",
      "ident": "NoiseReductionType",
      "$ref": "(resource) realtime > (model) noise_reduction_type > (schema)"
    },
    "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio/properties/input/properties/noise_reduction/anyOf/0/properties/type",
    "deprecated": false,
    "schemaType": "enum",
    "modelImplicit": false,
    "modelPath": "(resource) realtime > (model) noise_reduction_type",
    "childrenParentSchema": "enum",
    "children": [
      "(resource) realtime > (model) noise_reduction_type > (schema) > (member) 0",
      "(resource) realtime > (model) noise_reduction_type > (schema) > (member) 1"
    ]
  },
  "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) transcription > (property) model": {
    "kind": "HttpDeclProperty",
    "docstring": "The transcription model to use for source transcript deltas.",
    "key": "model",
    "optional": false,
    "nullable": false,
    "type": {
      "kind": "HttpTypeString"
    },
    "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio/properties/input/properties/transcription/anyOf/0/properties/model",
    "deprecated": false,
    "schemaType": "string",
    "children": []
  },
  "(resource) realtime > (model) noise_reduction_type > (schema) > (member) 0": {
    "kind": "HttpDeclReference",
    "type": {
      "kind": "HttpTypeLiteral",
      "literal": "near_field"
    }
  },
  "(resource) realtime > (model) noise_reduction_type > (schema) > (member) 1": {
    "kind": "HttpDeclReference",
    "type": {
      "kind": "HttpTypeLiteral",
      "literal": "far_field"
    }
  },
  "(resource) realtime > (model) noise_reduction_type > (schema)": {
    "kind": "HttpDeclTypeAlias",
    "oasRef": "#/components/schemas/NoiseReductionType",
    "ident": "NoiseReductionType",
    "type": {
      "kind": "HttpTypeUnion",
      "types": [
        {
          "kind": "HttpTypeLiteral",
          "literal": "near_field"
        },
        {
          "kind": "HttpTypeLiteral",
          "literal": "far_field"
        }
      ],
      "oasRef": "#/components/schemas/NoiseReductionType"
    },
    "docstring": "Type of noise reduction. `near_field` is for close-talking microphones such as headphones, `far_field` is for far-field microphones such as laptop or conference room microphones.\n",
    "childrenParentSchema": "enum",
    "children": [
      "(resource) realtime > (model) noise_reduction_type > (schema) > (member) 0",
      "(resource) realtime > (model) noise_reduction_type > (schema) > (member) 1"
    ]
  }
}

Example

{
  "type": "session.update",
  "session": {
    "audio": {
      "input": {
        "transcription": {
          "model": "gpt-realtime-whisper"
        },
        "noise_reduction": null
      },
      "output": {
        "language": "es"
      }
    }
  }
}

session.input_audio_buffer.append

Send this event to append audio bytes to the translation session input audio buffer.

WebSocket translation sessions accept base64-encoded 24 kHz PCM16 mono little-endian raw audio bytes. Unsupported websocket audio formats return a validation error because lower-quality audio materially degrades translation quality.

Translation consumes 200 ms engine frames. For best realtime behavior, append audio in 200 ms chunks. If a chunk is shorter, the server buffers it until it has enough audio for one frame. If a chunk is longer, the server splits it into 200 ms frames and enqueues them back-to-back.

Keep appending silence while the session is active. If a client stops sending audio and later resumes, model time treats the resumed audio as contiguous with the previous audio rather than as a real-world pause.

Schema

Schema name: RealtimeTranslationClientEventInputAudioBufferAppend

{
  "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema)": {
    "kind": "HttpDeclTypeAlias",
    "oasRef": "#/components/schemas/RealtimeTranslationClientEventInputAudioBufferAppend",
    "ident": "RealtimeTranslationInputAudioBufferAppendEvent",
    "type": {
      "kind": "HttpTypeObject",
      "members": [
        {
          "ident": "audio"
        },
        {
          "ident": "type"
        },
        {
          "ident": "event_id"
        }
      ]
    },
    "docstring": "Send this event to append audio bytes to the translation session input audio buffer.\n\nWebSocket translation sessions accept base64-encoded 24 kHz PCM16 mono\nlittle-endian raw audio bytes. Unsupported websocket audio formats return a\nvalidation error because lower-quality audio materially degrades translation\nquality.\n\nTranslation consumes 200 ms engine frames. For best realtime behavior, append\naudio in 200 ms chunks. If a chunk is shorter, the server buffers it until it\nhas enough audio for one frame. If a chunk is longer, the server splits it into\n200 ms frames and enqueues them back-to-back.\n\nKeep appending silence while the session is active. If a client stops sending\naudio and later resumes, model time treats the resumed audio as contiguous with\nthe previous audio rather than as a real-world pause.\n",
    "childrenParentSchema": "object",
    "children": [
      "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) audio",
      "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) type",
      "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) event_id"
    ]
  },
  "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) audio": {
    "kind": "HttpDeclProperty",
    "docstring": "Base64-encoded 24 kHz PCM16 mono audio bytes.",
    "key": "audio",
    "optional": false,
    "nullable": false,
    "type": {
      "kind": "HttpTypeString"
    },
    "oasRef": "#/components/schemas/RealtimeTranslationClientEventInputAudioBufferAppend/properties/audio",
    "deprecated": false,
    "schemaType": "string",
    "children": []
  },
  "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) type": {
    "kind": "HttpDeclProperty",
    "docstring": "The event type, must be `session.input_audio_buffer.append`.",
    "key": "type",
    "optional": false,
    "nullable": false,
    "type": {
      "kind": "HttpTypeUnion",
      "types": [
        {
          "kind": "HttpTypeLiteral",
          "literal": "session.input_audio_buffer.append"
        }
      ],
      "oasRef": "#/components/schemas/RealtimeTranslationClientEventInputAudioBufferAppend/properties/type"
    },
    "oasRef": "#/components/schemas/RealtimeTranslationClientEventInputAudioBufferAppend/properties/type",
    "deprecated": false,
    "schemaType": "enum",
    "childrenParentSchema": "enum",
    "children": [
      "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) type > (member) 0"
    ]
  },
  "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) event_id": {
    "kind": "HttpDeclProperty",
    "docstring": "Optional client-generated ID used to identify this event.",
    "key": "event_id",
    "optional": true,
    "nullable": false,
    "type": {
      "kind": "HttpTypeString"
    },
    "constraints": {
      "maxLength": 512
    },
    "oasRef": "#/components/schemas/RealtimeTranslationClientEventInputAudioBufferAppend/properties/event_id",
    "deprecated": false,
    "schemaType": "string",
    "children": []
  },
  "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) type > (member) 0": {
    "kind": "HttpDeclReference",
    "type": {
      "kind": "HttpTypeLiteral",
      "literal": "session.input_audio_buffer.append"
    }
  }
}

Example

{
  "event_id": "event_456",
  "type": "session.input_audio_buffer.append",
  "audio": "Base64EncodedAudioData"
}

session.close

Gracefully close the realtime translation session. The server flushes pending input audio and emits any remaining translated output before closing the session.

Schema

Schema name: RealtimeTranslationClientEventSessionClose

{
  "(resource) realtime > (model) realtime_translation_session_close_event > (schema)": {
    "kind": "HttpDeclTypeAlias",
    "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionClose",
    "ident": "RealtimeTranslationSessionCloseEvent",
    "type": {
      "kind": "HttpTypeObject",
      "members": [
        {
          "ident": "type"
        },
        {
          "ident": "event_id"
        }
      ]
    },
    "docstring": "Gracefully close the realtime translation session. The server flushes pending\ninput audio and emits any remaining translated output before closing the\nsession.\n",
    "childrenParentSchema": "object",
    "children": [
      "(resource) realtime > (model) realtime_translation_session_close_event > (schema) > (property) type",
      "(resource) realtime > (model) realtime_translation_session_close_event > (schema) > (property) event_id"
    ]
  },
  "(resource) realtime > (model) realtime_translation_session_close_event > (schema) > (property) type": {
    "kind": "HttpDeclProperty",
    "docstring": "The event type, must be `session.close`.",
    "key": "type",
    "optional": false,
    "nullable": false,
    "type": {
      "kind": "HttpTypeUnion",
      "types": [
        {
          "kind": "HttpTypeLiteral",
          "literal": "session.close"
        }
      ],
      "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionClose/properties/type"
    },
    "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionClose/properties/type",
    "deprecated": false,
    "schemaType": "enum",
    "childrenParentSchema": "enum",
    "children": [
      "(resource) realtime > (model) realtime_translation_session_close_event > (schema) > (property) type > (member) 0"
    ]
  },
  "(resource) realtime > (model) realtime_translation_session_close_event > (schema) > (property) event_id": {
    "kind": "HttpDeclProperty",
    "docstring": "Optional client-generated ID used to identify this event.",
    "key": "event_id",
    "optional": true,
    "nullable": false,
    "type": {
      "kind": "HttpTypeString"
    },
    "constraints": {
      "maxLength": 512
    },
    "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionClose/properties/event_id",
    "deprecated": false,
    "schemaType": "string",
    "children": []
  },
  "(resource) realtime > (model) realtime_translation_session_close_event > (schema) > (property) type > (member) 0": {
    "kind": "HttpDeclReference",
    "type": {
      "kind": "HttpTypeLiteral",
      "literal": "session.close"
    }
  }
}

Example

{
  "event_id": "event_789",
  "type": "session.close"
}

resources/realtime/translation-client-events.md +577 −0 created

1# Realtime translation client events

3These are events that the OpenAI Realtime Translation WebSocket server will accept from the client.

5## session.update

7Send this event to update the translation session configuration. Translation

8sessions support updates to `audio.output.language`, `audio.input.transcription`,

9and `audio.input.noise_reduction`.

11### Schema

13Schema name: `RealtimeTranslationClientEventSessionUpdate`

15```json

16{

17 "(resource) realtime > (model) realtime_translation_session_update_event > (schema)": {

18 "kind": "HttpDeclTypeAlias",

19 "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionUpdate",

20 "ident": "RealtimeTranslationSessionUpdateEvent",

21 "type": {

22 "kind": "HttpTypeObject",

23 "members": [

24 {

25 "ident": "session"

26 },

27 {

28 "ident": "type"

29 },

30 {

31 "ident": "event_id"

32 }

33 ]

34 },

35 "docstring": "Send this event to update the translation session configuration. Translation\nsessions support updates to `audio.output.language`, `audio.input.transcription`,\nand `audio.input.noise_reduction`.\n",

36 "childrenParentSchema": "object",

37 "children": [

38 "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) session",

39 "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) type",

40 "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) event_id"

41 ]

42 },

43 "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) session": {

44 "kind": "HttpDeclProperty",

45 "title": "Realtime translation session update",

46 "docstring": "Translation session fields to update. The session `type` and `model` are set\nat creation and cannot be changed with `session.update`.\n",

47 "key": "session",

48 "optional": false,

49 "nullable": false,

50 "type": {

51 "kind": "HttpTypeReference",

52 "ident": "RealtimeTranslationSessionUpdateRequest",

53 "$ref": "(resource) realtime > (model) realtime_translation_session_update_request > (schema)"

54 },

55 "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionUpdate/properties/session",

56 "deprecated": false,

57 "schemaType": "object",

58 "modelImplicit": false,

59 "modelPath": "(resource) realtime > (model) realtime_translation_session_update_request",

60 "childrenParentSchema": "object",

61 "children": [

62 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio"

63 ]

64 },

65 "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) type": {

66 "kind": "HttpDeclProperty",

67 "docstring": "The event type, must be `session.update`.",

68 "key": "type",

69 "optional": false,

70 "nullable": false,

71 "type": {

72 "kind": "HttpTypeUnion",

73 "types": [

74 {

75 "kind": "HttpTypeLiteral",

76 "literal": "session.update"

77 }

78 ],

79 "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionUpdate/properties/type"

80 },

81 "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionUpdate/properties/type",

82 "deprecated": false,

83 "schemaType": "enum",

84 "childrenParentSchema": "enum",

85 "children": [

86 "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) type > (member) 0"

87 ]

88 },

89 "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) event_id": {

90 "kind": "HttpDeclProperty",

91 "docstring": "Optional client-generated ID used to identify this event.",

92 "key": "event_id",

93 "optional": true,

94 "nullable": false,

95 "type": {

96 "kind": "HttpTypeString"

97 },

98 "constraints": {

99 "maxLength": 512

100 },

101 "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionUpdate/properties/event_id",

102 "deprecated": false,

103 "schemaType": "string",

104 "children": []

105 },

106 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio": {

107 "kind": "HttpDeclProperty",

108 "docstring": "Configuration for translation input and output audio.\n",

109 "key": "audio",

110 "optional": true,

111 "nullable": false,

112 "type": {

113 "kind": "HttpTypeObject",

114 "members": [

115 {

116 "ident": "input"

117 },

118 {

119 "ident": "output"

120 }

121 ]

122 },

123 "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio",

124 "deprecated": false,

125 "schemaType": "object",

126 "childrenParentSchema": "object",

127 "children": [

128 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input",

129 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) output"

130 ]

131 },

132 "(resource) realtime > (model) realtime_translation_session_update_request > (schema)": {

133 "kind": "HttpDeclTypeAlias",

134 "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest",

135 "ident": "RealtimeTranslationSessionUpdateRequest",

136 "type": {

137 "kind": "HttpTypeObject",

138 "members": [

139 {

140 "ident": "audio"

141 }

142 ]

143 },

144 "docstring": "Realtime translation session fields that can be updated with `session.update`.\n",

145 "childrenParentSchema": "object",

146 "children": [

147 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio"

148 ]

149 },

150 "(resource) realtime > (model) realtime_translation_session_update_event > (schema) > (property) type > (member) 0": {

151 "kind": "HttpDeclReference",

152 "type": {

153 "kind": "HttpTypeLiteral",

154 "literal": "session.update"

155 }

156 },

157 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input": {

158 "kind": "HttpDeclProperty",

159 "key": "input",

160 "optional": true,

161 "nullable": false,

162 "type": {

163 "kind": "HttpTypeObject",

164 "members": [

165 {

166 "ident": "noise_reduction"

167 },

168 {

169 "ident": "transcription"

170 }

171 ]

172 },

173 "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio/properties/input",

174 "deprecated": false,

175 "schemaType": "object",

176 "childrenParentSchema": "object",

177 "children": [

178 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) noise_reduction",

179 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) transcription"

180 ]

181 },

182 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) output": {

183 "kind": "HttpDeclProperty",

184 "key": "output",

185 "optional": true,

186 "nullable": false,

187 "type": {

188 "kind": "HttpTypeObject",

189 "members": [

190 {

191 "ident": "language"

192 }

193 ]

194 },

195 "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio/properties/output",

196 "deprecated": false,

197 "schemaType": "object",

198 "childrenParentSchema": "object",

199 "children": [

200 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) output > (property) language"

201 ]

202 },

203 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) noise_reduction": {

204 "kind": "HttpDeclProperty",

205 "docstring": "Optional input noise reduction. Set to `null` to disable it.\n",

206 "key": "noise_reduction",

207 "optional": true,

208 "nullable": true,

209 "type": {

210 "kind": "HttpTypeObject",

211 "members": [

212 {

213 "ident": "type"

214 }

215 ]

216 },

217 "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio/properties/input/properties/noise_reduction",

218 "deprecated": false,

219 "schemaType": "object",

220 "childrenParentSchema": "object",

221 "children": [

222 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) noise_reduction > (property) type"

223 ]

224 },

225 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) transcription": {

226 "kind": "HttpDeclProperty",

227 "docstring": "Optional source-language transcription. When configured, the server emits\n`session.input_transcript.delta` events. Translation itself still runs from\nthe input audio stream.\n",

228 "key": "transcription",

229 "optional": true,

230 "nullable": true,

231 "type": {

232 "kind": "HttpTypeObject",

233 "members": [

234 {

235 "ident": "model"

236 }

237 ]

238 },

239 "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio/properties/input/properties/transcription",

240 "deprecated": false,

241 "schemaType": "object",

242 "childrenParentSchema": "object",

243 "children": [

244 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) transcription > (property) model"

245 ]

246 },

247 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) output > (property) language": {

248 "kind": "HttpDeclProperty",

249 "docstring": "Target language for translated output audio and transcript deltas.\n",

250 "key": "language",

251 "optional": true,

252 "nullable": false,

253 "type": {

254 "kind": "HttpTypeString"

255 },

256 "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio/properties/output/properties/language",

257 "deprecated": false,

258 "schemaType": "string",

259 "children": []

260 },

261 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) noise_reduction > (property) type": {

262 "kind": "HttpDeclProperty",

263 "docstring": "Type of noise reduction. `near_field` is for close-talking microphones such as headphones, `far_field` is for far-field microphones such as laptop or conference room microphones.\n",

264 "key": "type",

265 "optional": false,

266 "nullable": false,

267 "type": {

268 "kind": "HttpTypeReference",

269 "ident": "NoiseReductionType",

270 "$ref": "(resource) realtime > (model) noise_reduction_type > (schema)"

271 },

272 "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio/properties/input/properties/noise_reduction/anyOf/0/properties/type",

273 "deprecated": false,

274 "schemaType": "enum",

275 "modelImplicit": false,

276 "modelPath": "(resource) realtime > (model) noise_reduction_type",

277 "childrenParentSchema": "enum",

278 "children": [

279 "(resource) realtime > (model) noise_reduction_type > (schema) > (member) 0",

280 "(resource) realtime > (model) noise_reduction_type > (schema) > (member) 1"

281 ]

282 },

283 "(resource) realtime > (model) realtime_translation_session_update_request > (schema) > (property) audio > (property) input > (property) transcription > (property) model": {

284 "kind": "HttpDeclProperty",

285 "docstring": "The transcription model to use for source transcript deltas.",

286 "key": "model",

287 "optional": false,

288 "nullable": false,

289 "type": {

290 "kind": "HttpTypeString"

291 },

292 "oasRef": "#/components/schemas/RealtimeTranslationSessionUpdateRequest/properties/audio/properties/input/properties/transcription/anyOf/0/properties/model",

293 "deprecated": false,

294 "schemaType": "string",

295 "children": []

296 },

297 "(resource) realtime > (model) noise_reduction_type > (schema) > (member) 0": {

298 "kind": "HttpDeclReference",

299 "type": {

300 "kind": "HttpTypeLiteral",

301 "literal": "near_field"

302 }

303 },

304 "(resource) realtime > (model) noise_reduction_type > (schema) > (member) 1": {

305 "kind": "HttpDeclReference",

306 "type": {

307 "kind": "HttpTypeLiteral",

308 "literal": "far_field"

309 }

310 },

311 "(resource) realtime > (model) noise_reduction_type > (schema)": {

312 "kind": "HttpDeclTypeAlias",

313 "oasRef": "#/components/schemas/NoiseReductionType",

314 "ident": "NoiseReductionType",

315 "type": {

316 "kind": "HttpTypeUnion",

317 "types": [

318 {

319 "kind": "HttpTypeLiteral",

320 "literal": "near_field"

321 },

322 {

323 "kind": "HttpTypeLiteral",

324 "literal": "far_field"

325 }

326 ],

327 "oasRef": "#/components/schemas/NoiseReductionType"

328 },

329 "docstring": "Type of noise reduction. `near_field` is for close-talking microphones such as headphones, `far_field` is for far-field microphones such as laptop or conference room microphones.\n",

330 "childrenParentSchema": "enum",

331 "children": [

332 "(resource) realtime > (model) noise_reduction_type > (schema) > (member) 0",

333 "(resource) realtime > (model) noise_reduction_type > (schema) > (member) 1"

334 ]

335 }

336}

337```

338

339### Example

340

341```json

342{

343 "type": "session.update",

344 "session": {

345 "audio": {

346 "input": {

347 "transcription": {

348 "model": "gpt-realtime-whisper"

349 },

350 "noise_reduction": null

351 },

352 "output": {

353 "language": "es"

354 }

355 }

356 }

357}

358```

359

360## session.input_audio_buffer.append

361

362Send this event to append audio bytes to the translation session input audio buffer.

363

364WebSocket translation sessions accept base64-encoded 24 kHz PCM16 mono

365little-endian raw audio bytes. Unsupported websocket audio formats return a

366validation error because lower-quality audio materially degrades translation

367quality.

368

369Translation consumes 200 ms engine frames. For best realtime behavior, append

370audio in 200 ms chunks. If a chunk is shorter, the server buffers it until it

371has enough audio for one frame. If a chunk is longer, the server splits it into

372200 ms frames and enqueues them back-to-back.

373

374Keep appending silence while the session is active. If a client stops sending

375audio and later resumes, model time treats the resumed audio as contiguous with

376the previous audio rather than as a real-world pause.

377

378### Schema

379

380Schema name: `RealtimeTranslationClientEventInputAudioBufferAppend`

381

382```json

383{

384 "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema)": {

385 "kind": "HttpDeclTypeAlias",

386 "oasRef": "#/components/schemas/RealtimeTranslationClientEventInputAudioBufferAppend",

387 "ident": "RealtimeTranslationInputAudioBufferAppendEvent",

388 "type": {

389 "kind": "HttpTypeObject",

390 "members": [

391 {

392 "ident": "audio"

393 },

394 {

395 "ident": "type"

396 },

397 {

398 "ident": "event_id"

399 }

400 ]

401 },

402 "docstring": "Send this event to append audio bytes to the translation session input audio buffer.\n\nWebSocket translation sessions accept base64-encoded 24 kHz PCM16 mono\nlittle-endian raw audio bytes. Unsupported websocket audio formats return a\nvalidation error because lower-quality audio materially degrades translation\nquality.\n\nTranslation consumes 200 ms engine frames. For best realtime behavior, append\naudio in 200 ms chunks. If a chunk is shorter, the server buffers it until it\nhas enough audio for one frame. If a chunk is longer, the server splits it into\n200 ms frames and enqueues them back-to-back.\n\nKeep appending silence while the session is active. If a client stops sending\naudio and later resumes, model time treats the resumed audio as contiguous with\nthe previous audio rather than as a real-world pause.\n",

403 "childrenParentSchema": "object",

404 "children": [

405 "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) audio",

406 "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) type",

407 "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) event_id"

408 ]

409 },

410 "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) audio": {

411 "kind": "HttpDeclProperty",

412 "docstring": "Base64-encoded 24 kHz PCM16 mono audio bytes.",

413 "key": "audio",

414 "optional": false,

415 "nullable": false,

416 "type": {

417 "kind": "HttpTypeString"

418 },

419 "oasRef": "#/components/schemas/RealtimeTranslationClientEventInputAudioBufferAppend/properties/audio",

420 "deprecated": false,

421 "schemaType": "string",

422 "children": []

423 },

424 "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) type": {

425 "kind": "HttpDeclProperty",

426 "docstring": "The event type, must be `session.input_audio_buffer.append`.",

427 "key": "type",

428 "optional": false,

429 "nullable": false,

430 "type": {

431 "kind": "HttpTypeUnion",

432 "types": [

433 {

434 "kind": "HttpTypeLiteral",

435 "literal": "session.input_audio_buffer.append"

436 }

437 ],

438 "oasRef": "#/components/schemas/RealtimeTranslationClientEventInputAudioBufferAppend/properties/type"

439 },

440 "oasRef": "#/components/schemas/RealtimeTranslationClientEventInputAudioBufferAppend/properties/type",

441 "deprecated": false,

442 "schemaType": "enum",

443 "childrenParentSchema": "enum",

444 "children": [

445 "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) type > (member) 0"

446 ]

447 },

448 "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) event_id": {

449 "kind": "HttpDeclProperty",

450 "docstring": "Optional client-generated ID used to identify this event.",

451 "key": "event_id",

452 "optional": true,

453 "nullable": false,

454 "type": {

455 "kind": "HttpTypeString"

456 },

457 "constraints": {

458 "maxLength": 512

459 },

460 "oasRef": "#/components/schemas/RealtimeTranslationClientEventInputAudioBufferAppend/properties/event_id",

461 "deprecated": false,

462 "schemaType": "string",

463 "children": []

464 },

465 "(resource) realtime > (model) realtime_translation_input_audio_buffer_append_event > (schema) > (property) type > (member) 0": {

466 "kind": "HttpDeclReference",

467 "type": {

468 "kind": "HttpTypeLiteral",

469 "literal": "session.input_audio_buffer.append"

470 }

471 }

472}

473```

474

475### Example

476

477```json

478{

479 "event_id": "event_456",

480 "type": "session.input_audio_buffer.append",

481 "audio": "Base64EncodedAudioData"

482}

483```

484

485## session.close

486

487Gracefully close the realtime translation session. The server flushes pending

488input audio and emits any remaining translated output before closing the

489session.

490

491### Schema

492

493Schema name: `RealtimeTranslationClientEventSessionClose`

494

495```json

496{

497 "(resource) realtime > (model) realtime_translation_session_close_event > (schema)": {

498 "kind": "HttpDeclTypeAlias",

499 "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionClose",

500 "ident": "RealtimeTranslationSessionCloseEvent",

501 "type": {

502 "kind": "HttpTypeObject",

503 "members": [

504 {

505 "ident": "type"

506 },

507 {

508 "ident": "event_id"

509 }

510 ]

511 },

512 "docstring": "Gracefully close the realtime translation session. The server flushes pending\ninput audio and emits any remaining translated output before closing the\nsession.\n",

513 "childrenParentSchema": "object",

514 "children": [

515 "(resource) realtime > (model) realtime_translation_session_close_event > (schema) > (property) type",

516 "(resource) realtime > (model) realtime_translation_session_close_event > (schema) > (property) event_id"

517 ]

518 },

519 "(resource) realtime > (model) realtime_translation_session_close_event > (schema) > (property) type": {

520 "kind": "HttpDeclProperty",

521 "docstring": "The event type, must be `session.close`.",

522 "key": "type",

523 "optional": false,

524 "nullable": false,

525 "type": {

526 "kind": "HttpTypeUnion",

527 "types": [

528 {

529 "kind": "HttpTypeLiteral",

530 "literal": "session.close"

531 }

532 ],

533 "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionClose/properties/type"

534 },

535 "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionClose/properties/type",

536 "deprecated": false,

537 "schemaType": "enum",

538 "childrenParentSchema": "enum",

539 "children": [

540 "(resource) realtime > (model) realtime_translation_session_close_event > (schema) > (property) type > (member) 0"

541 ]

542 },

543 "(resource) realtime > (model) realtime_translation_session_close_event > (schema) > (property) event_id": {

544 "kind": "HttpDeclProperty",

545 "docstring": "Optional client-generated ID used to identify this event.",

546 "key": "event_id",

547 "optional": true,

548 "nullable": false,

549 "type": {

550 "kind": "HttpTypeString"

551 },

552 "constraints": {

553 "maxLength": 512

554 },

555 "oasRef": "#/components/schemas/RealtimeTranslationClientEventSessionClose/properties/event_id",

556 "deprecated": false,

557 "schemaType": "string",

558 "children": []

559 },

560 "(resource) realtime > (model) realtime_translation_session_close_event > (schema) > (property) type > (member) 0": {

561 "kind": "HttpDeclReference",

562 "type": {

563 "kind": "HttpTypeLiteral",

564 "literal": "session.close"

565 }

566 }

567}

568```

569

570### Example

571

572```json

573{

574 "event_id": "event_789",

575 "type": "session.close"

576}

577```

resources/realtime/translation-client-events.md 2026-07-07 08:02 UTC to 2026-07-09 20:58 UTC

Realtime translation client events

session.update

Schema

Example

session.input_audio_buffer.append

Schema

Example

session.close

Schema

Example