{
  "openapi": "3.1.0",
  "info": {
    "title": "kiapi Speech API",
    "description": "Text-to-speech in a registered voice (OpenAI `audio/speech` compatible).\n\nRegister a voice once from a few seconds of reference audio and its transcript\n(`POST /v1/audio/voices`), then pass its id as `voice` to `POST /v1/audio/speech`.\nThe speech clones that voice: there are no built-in voices.\n\n`response_format: pcm` (default) streams raw 24 kHz, 16-bit, mono, little-endian\nPCM while it is generated, so playback can start on the first chunk (about half a\nsecond on Apple M1 Max). `wav` returns one file when generation finishes. Closing\nthe connection stops the generation.\n\nRequests share kiapi's single worker with every other capability: while another\njob runs, speech waits for it.\n\n## Upstream docs\n- [Qwen3-TTS](https://github.com/QwenLM/Qwen3-TTS) — the model family\n- [mlx-audio](https://github.com/Blaizzy/mlx-audio) — the MLX implementation\n\n## Models\n- **qwen3-tts-1.7b** (default, alias `qwen3-tts`) —\n  `mlx-community/Qwen3-TTS-12Hz-1.7B-Base-bf16`, Apache-2.0. Speaks Chinese,\n  English, Japanese, Korean, German, French, Russian, Portuguese, Spanish, and\n  Italian.\n\nDiscover models at `GET /v1/audio/speech/models`.\n\n## Voice Tips\n- Use 5 to 15 seconds of one clear speaker, without music or noise.\n- Give the transcript word for word, or omit it to have kiapi transcribe the\n  reference (`KIAPI_SPEECH_TRANSCRIBE_REFERENCE`). Without any transcript the\n  voice is cloned from the speaker embedding only and sounds less like the\n  reference.\n- Only clone voices you have the right to use.\n",
    "version": "0.8.0"
  },
  "paths": {
    "/v1/audio/speech": {
      "post": {
        "summary": "Create Speech",
        "description": "Speak `input` in a registered voice (OpenAI `audio/speech` compatible).\n\n`voice` is the id of a voice registered with POST /v1/audio/voices; there are\nno built-in voices. With `response_format: pcm` (default) the audio streams as\nit is generated, and closing the connection stops the generation. With `wav`\nthe whole file is returned when generation finishes.\n\nThe request waits in kiapi's single worker queue while another job runs.",
        "operationId": "create_speech_v1_audio_speech_post",
        "parameters": [
          {
            "name": "Accept",
            "in": "header",
            "required": false,
            "schema": {
              "anyOf": [
                {
                  "type": "string"
                },
                {
                  "type": "null"
                }
              ],
              "description": "Response media type preference. application/json returns the Job JSON; otherwise sync requests with one artifact return raw bytes when possible.",
              "examples": [
                "application/json",
                "image/png",
                "audio/wav",
                "video/mp4"
              ],
              "title": "Accept"
            },
            "description": "Response media type preference. application/json returns the Job JSON; otherwise sync requests with one artifact return raw bytes when possible."
          }
        ],
        "requestBody": {
          "required": true,
          "content": {
            "application/json": {
              "schema": {
                "$ref": "#/components/schemas/SpeechRequest"
              }
            }
          }
        },
        "responses": {
          "200": {
            "description": "Sync result. Returns Job JSON with Accept: application/json; single-artifact jobs may return raw bytes otherwise.",
            "content": {
              "application/json": {
                "schema": {
                  "$ref": "#/components/schemas/JobSpeechResponse"
                }
              },
              "audio/wav": {
                "schema": {
                  "type": "string",
                  "format": "binary"
                }
              },
              "audio/pcm": {
                "schema": {
                  "type": "string",
                  "format": "binary"
                }
              }
            },
            "headers": {
              "X-Kiapi-File-Id": {
                "description": "Produced artifact file_id when raw bytes are returned.",
                "schema": {
                  "type": "string"
                }
              },
              "X-Kiapi-Job-Id": {
                "description": "Job id when raw bytes are returned.",
                "schema": {
                  "type": "string"
                }
              }
            }
          },
          "400": {
            "description": "Invalid request for the selected model or file reference."
          },
          "422": {
            "description": "Request schema or validation error."
          },
          "503": {
            "description": "Model setup or memory budget error."
          },
          "504": {
            "description": "Sync request exceeded the configured timeout."
          }
        }
      }
    },
    "/v1/audio/voices": {
      "get": {
        "summary": "List Voices",
        "description": "List registered voices, oldest first.",
        "operationId": "list_voices_v1_audio_voices_get",
        "responses": {
          "200": {
            "description": "Successful Response",
            "content": {
              "application/json": {
                "schema": {
                  "$ref": "#/components/schemas/VoiceList"
                }
              }
            }
          }
        }
      },
      "post": {
        "summary": "Create Voice",
        "description": "Register a voice from reference audio and its transcript.\n\nWithout `transcript`, kiapi transcribes the reference (after cutting it to\nthe maximum length) with the transcriptions family and stores that text. The\ntranscription waits in kiapi's single worker queue like any other job.",
        "operationId": "create_voice_v1_audio_voices_post",
        "requestBody": {
          "content": {
            "application/json": {
              "schema": {
                "$ref": "#/components/schemas/VoiceCreateRequest"
              }
            }
          },
          "required": true
        },
        "responses": {
          "200": {
            "description": "Successful Response",
            "content": {
              "application/json": {
                "schema": {
                  "$ref": "#/components/schemas/Voice"
                }
              }
            }
          },
          "422": {
            "description": "Validation Error",
            "content": {
              "application/json": {
                "schema": {
                  "$ref": "#/components/schemas/HTTPValidationError"
                }
              }
            }
          }
        }
      }
    },
    "/v1/audio/voices/{voice_id}": {
      "get": {
        "summary": "Get Voice",
        "operationId": "get_voice_v1_audio_voices__voice_id__get",
        "parameters": [
          {
            "name": "voice_id",
            "in": "path",
            "required": true,
            "schema": {
              "type": "string",
              "title": "Voice Id"
            }
          }
        ],
        "responses": {
          "200": {
            "description": "Successful Response",
            "content": {
              "application/json": {
                "schema": {
                  "$ref": "#/components/schemas/Voice"
                }
              }
            }
          },
          "422": {
            "description": "Validation Error",
            "content": {
              "application/json": {
                "schema": {
                  "$ref": "#/components/schemas/HTTPValidationError"
                }
              }
            }
          }
        }
      },
      "patch": {
        "summary": "Update Voice",
        "description": "Change a voice's name or transcript; omitted fields stay as they are.",
        "operationId": "update_voice_v1_audio_voices__voice_id__patch",
        "parameters": [
          {
            "name": "voice_id",
            "in": "path",
            "required": true,
            "schema": {
              "type": "string",
              "title": "Voice Id"
            }
          }
        ],
        "requestBody": {
          "required": true,
          "content": {
            "application/json": {
              "schema": {
                "$ref": "#/components/schemas/VoiceUpdateRequest"
              }
            }
          }
        },
        "responses": {
          "200": {
            "description": "Successful Response",
            "content": {
              "application/json": {
                "schema": {
                  "$ref": "#/components/schemas/Voice"
                }
              }
            }
          },
          "422": {
            "description": "Validation Error",
            "content": {
              "application/json": {
                "schema": {
                  "$ref": "#/components/schemas/HTTPValidationError"
                }
              }
            }
          }
        }
      },
      "delete": {
        "summary": "Delete Voice",
        "operationId": "delete_voice_v1_audio_voices__voice_id__delete",
        "parameters": [
          {
            "name": "voice_id",
            "in": "path",
            "required": true,
            "schema": {
              "type": "string",
              "title": "Voice Id"
            }
          }
        ],
        "responses": {
          "200": {
            "description": "Successful Response",
            "content": {
              "application/json": {
                "schema": {
                  "type": "object",
                  "additionalProperties": true,
                  "title": "Response Delete Voice V1 Audio Voices  Voice Id  Delete"
                }
              }
            }
          },
          "422": {
            "description": "Validation Error",
            "content": {
              "application/json": {
                "schema": {
                  "$ref": "#/components/schemas/HTTPValidationError"
                }
              }
            }
          }
        }
      }
    },
    "/v1/audio/voices/{voice_id}/audio": {
      "get": {
        "summary": "Get Voice Audio",
        "description": "The voice's reference audio as stored: 24 kHz, 16-bit, mono WAV.",
        "operationId": "get_voice_audio_v1_audio_voices__voice_id__audio_get",
        "parameters": [
          {
            "name": "voice_id",
            "in": "path",
            "required": true,
            "schema": {
              "type": "string",
              "title": "Voice Id"
            }
          }
        ],
        "responses": {
          "200": {
            "description": "Successful Response",
            "content": {
              "application/json": {
                "schema": {}
              },
              "audio/wav": {
                "schema": {
                  "type": "string",
                  "format": "binary"
                }
              }
            }
          },
          "422": {
            "description": "Validation Error",
            "content": {
              "application/json": {
                "schema": {
                  "$ref": "#/components/schemas/HTTPValidationError"
                }
              }
            }
          }
        }
      }
    },
    "/v1/audio/speech/models": {
      "get": {
        "summary": "List Models",
        "description": "List the servable models for this capability.\n\nReturns the public catalog of every variant selectable via the ``model``\nfield on this capability's endpoints.",
        "operationId": "list_models_v1_audio_speech_models_get",
        "responses": {
          "200": {
            "description": "Successful Response",
            "content": {
              "application/json": {
                "schema": {
                  "items": {
                    "$ref": "#/components/schemas/CapabilityModelSpec"
                  },
                  "type": "array",
                  "title": "Response List Models V1 Audio Speech Models Get"
                }
              }
            }
          }
        }
      }
    }
  },
  "components": {
    "schemas": {
      "CapabilityModelSpec": {
        "properties": {
          "name": {
            "type": "string",
            "title": "Name",
            "description": "Model variant name to pass in the request model field.",
            "examples": [
              "turbo"
            ]
          },
          "family": {
            "type": "string",
            "title": "Family",
            "description": "Capability family that resolves this model variant.",
            "examples": [
              "zimage"
            ]
          },
          "domain": {
            "type": "string",
            "title": "Domain",
            "description": "Capability domain used for grouping model lists.",
            "examples": [
              "image"
            ]
          },
          "aliases": {
            "items": {
              "type": "string"
            },
            "type": "array",
            "title": "Aliases",
            "description": "Alternative names that also resolve to this model.",
            "examples": [
              [
                "omni",
                "qwen3-omni-30b"
              ]
            ]
          },
          "default": {
            "type": "boolean",
            "title": "Default",
            "description": "Whether this is the default model when the request omits model.",
            "default": false,
            "examples": [
              true
            ]
          },
          "features": {
            "items": {
              "type": "string"
            },
            "type": "array",
            "title": "Features",
            "description": "Handler-declared modalities and features supported by this model.",
            "examples": [
              [
                "text",
                "image"
              ]
            ]
          }
        },
        "type": "object",
        "required": [
          "name",
          "family",
          "domain"
        ],
        "title": "CapabilityModelSpec",
        "description": "Public model discovery entry for capability-specific model lists."
      },
      "FileDataURLRef": {
        "properties": {
          "type": {
            "type": "string",
            "const": "data_url",
            "title": "Type",
            "default": "data_url"
          },
          "data_url": {
            "type": "string",
            "minLength": 1,
            "title": "Data Url"
          }
        },
        "type": "object",
        "required": [
          "data_url"
        ],
        "title": "FileDataURLRef",
        "examples": [
          {
            "data_url": "data:image/png;base64,iVBORw0KGgo...",
            "type": "data_url"
          }
        ]
      },
      "FileID": {
        "type": "string"
      },
      "FileIDRef": {
        "properties": {
          "type": {
            "type": "string",
            "const": "file_id",
            "title": "Type",
            "default": "file_id"
          },
          "file_id": {
            "type": "string",
            "minLength": 1,
            "title": "File Id"
          }
        },
        "type": "object",
        "required": [
          "file_id"
        ],
        "title": "FileIDRef",
        "examples": [
          {
            "file_id": "file_0123456789abcdef",
            "type": "file_id"
          }
        ]
      },
      "FileRef": {
        "oneOf": [
          {
            "$ref": "#/components/schemas/FileIDRef"
          },
          {
            "$ref": "#/components/schemas/FileURLRef"
          },
          {
            "$ref": "#/components/schemas/FileDataURLRef"
          }
        ],
        "discriminator": {
          "propertyName": "type",
          "mapping": {
            "data_url": "#/components/schemas/FileDataURLRef",
            "file_id": "#/components/schemas/FileIDRef",
            "url": "#/components/schemas/FileURLRef"
          }
        }
      },
      "FileURLRef": {
        "properties": {
          "type": {
            "type": "string",
            "const": "url",
            "title": "Type",
            "default": "url"
          },
          "url": {
            "type": "string",
            "minLength": 1,
            "title": "Url"
          }
        },
        "type": "object",
        "required": [
          "url"
        ],
        "title": "FileURLRef",
        "examples": [
          {
            "type": "url",
            "url": "https://example.com/input.png"
          }
        ]
      },
      "HTTPValidationError": {
        "properties": {
          "detail": {
            "items": {
              "$ref": "#/components/schemas/ValidationError"
            },
            "type": "array",
            "title": "Detail"
          }
        },
        "type": "object",
        "title": "HTTPValidationError"
      },
      "JobID": {
        "type": "string"
      },
      "JobSpeechResponse": {
        "properties": {
          "type": {
            "$ref": "#/components/schemas/JobType",
            "description": "Job type. Use this to interpret the capability-specific result payload.",
            "examples": [
              "zimage"
            ]
          },
          "params": {
            "additionalProperties": true,
            "type": "object",
            "title": "Params",
            "description": "Request parameters captured for inspection and reproducibility. Secret or large media payloads may be omitted or redacted by endpoints."
          },
          "id": {
            "$ref": "#/components/schemas/JobID",
            "description": "In-memory job id. Jobs are cleared when the kiapi process restarts.",
            "examples": [
              "job_0123456789abcdef"
            ]
          },
          "status": {
            "$ref": "#/components/schemas/JobStatus",
            "description": "Job lifecycle state: queued, running, succeeded, failed, or canceled.",
            "default": "queued",
            "examples": [
              "succeeded"
            ]
          },
          "result": {
            "anyOf": [
              {
                "$ref": "#/components/schemas/SpeechResponse"
              },
              {
                "type": "null"
              }
            ]
          },
          "artifacts": {
            "items": {
              "$ref": "#/components/schemas/FileID"
            },
            "type": "array",
            "title": "Artifacts",
            "description": "File ids produced by the job. Use GET /v1/files/{file_id} for metadata or /download for bytes.",
            "examples": [
              [
                "file_0123456789abcdef"
              ]
            ]
          },
          "error": {
            "anyOf": [
              {
                "type": "string"
              },
              {
                "type": "null"
              }
            ],
            "title": "Error",
            "description": "Error message when status is failed; otherwise null.",
            "examples": [
              "model 'turbo' is not activated"
            ]
          },
          "created_at": {
            "type": "number",
            "title": "Created At",
            "description": "Unix timestamp when the job was created.",
            "examples": [
              1766200000.0
            ]
          },
          "started_at": {
            "anyOf": [
              {
                "type": "number"
              },
              {
                "type": "null"
              }
            ],
            "title": "Started At",
            "description": "Unix timestamp when the worker started the job, or null while queued.",
            "examples": [
              1766200001.0
            ]
          },
          "finished_at": {
            "anyOf": [
              {
                "type": "number"
              },
              {
                "type": "null"
              }
            ],
            "title": "Finished At",
            "description": "Unix timestamp when the job reached a terminal state, or null while queued/running.",
            "examples": [
              1766200030.0
            ]
          },
          "progress": {
            "anyOf": [
              {
                "type": "number",
                "maximum": 1.0,
                "minimum": 0.0
              },
              {
                "type": "null"
              }
            ],
            "title": "Progress",
            "description": "Best-effort completion fraction in [0.0, 1.0]. Null means the job has not reported progress.",
            "examples": [
              0.42
            ]
          },
          "progress_label": {
            "type": "string",
            "title": "Progress Label",
            "description": "Short human-readable phase label such as queued, running, denoising, saving, or done.",
            "default": "queued",
            "examples": [
              "denoising"
            ]
          }
        },
        "type": "object",
        "required": [
          "type"
        ],
        "title": "JobSpeechResponse"
      },
      "JobStatus": {
        "type": "string",
        "enum": [
          "queued",
          "running",
          "succeeded",
          "failed",
          "canceled"
        ],
        "title": "JobStatus"
      },
      "JobType": {
        "type": "string"
      },
      "SpeechRequest": {
        "properties": {
          "model": {
            "anyOf": [
              {
                "type": "string"
              },
              {
                "type": "null"
              }
            ],
            "title": "Model",
            "description": "Speech model. Omit/null for the family default (`qwen3-tts-1.7b`). Discover models at GET /v1/audio/speech/models."
          },
          "input": {
            "type": "string",
            "minLength": 1,
            "title": "Input",
            "description": "Text to speak. Capped by `KIAPI_SPEECH_MAX_INPUT_CHARS` (default 4096). It is read as written, so leave out markup, lists, and URLs.",
            "examples": [
              "こんにちは。今日はいい天気ですね。"
            ]
          },
          "voice": {
            "type": "string",
            "minLength": 1,
            "title": "Voice",
            "description": "Id of a registered voice (POST /v1/audio/voices). There are no built-in voices: the speech clones the registered reference.",
            "examples": [
              "voice_0123456789abcdef"
            ]
          },
          "response_format": {
            "type": "string",
            "enum": [
              "pcm",
              "wav"
            ],
            "title": "Response format",
            "description": "`pcm` streams raw 24 kHz, 16-bit, mono, little-endian PCM while it is generated (the first chunk arrives before the rest is generated). `wav` returns one WAV file when generation finishes.",
            "default": "pcm"
          },
          "stream_format": {
            "type": "string",
            "const": "audio",
            "title": "Stream format",
            "description": "Only `audio` (raw bytes) is supported; `sse` is not.",
            "default": "audio"
          },
          "speed": {
            "type": "number",
            "title": "Speed",
            "description": "Only 1.0 is supported (Qwen3-TTS has no speed control).",
            "default": 1.0
          },
          "instructions": {
            "anyOf": [
              {
                "type": "string"
              },
              {
                "type": "null"
              }
            ],
            "title": "Instructions",
            "description": "Not supported by voice cloning; must be empty."
          },
          "language": {
            "anyOf": [
              {
                "type": "string"
              },
              {
                "type": "null"
              }
            ],
            "title": "Language",
            "description": "kiapi extension. Language of `input` as a BCP 47 tag or name (`ja`, `ja-JP`, `japanese`). Omit to let the model detect it. Qwen3-TTS speaks Chinese, English, Japanese, Korean, German, French, Russian, Portuguese, Spanish, and Italian.",
            "examples": [
              "ja"
            ]
          },
          "seed": {
            "anyOf": [
              {
                "type": "integer"
              },
              {
                "type": "null"
              }
            ],
            "title": "Seed",
            "description": "kiapi extension. Random seed; null picks one."
          }
        },
        "additionalProperties": true,
        "type": "object",
        "required": [
          "input",
          "voice"
        ],
        "title": "SpeechRequest",
        "description": "Text to speak in a registered voice (OpenAI `audio/speech` compatible)."
      },
      "SpeechResponse": {
        "properties": {
          "file_id": {
            "type": "string",
            "title": "File Id",
            "description": "Files-API id of the produced WAV."
          },
          "audio_bytes": {
            "type": "integer",
            "title": "Audio Bytes",
            "description": "Size of the produced WAV in bytes."
          },
          "model": {
            "type": "string",
            "title": "Model",
            "description": "Resolved model."
          },
          "voice": {
            "type": "string",
            "title": "Voice",
            "description": "Voice id that was cloned."
          },
          "language": {
            "type": "string",
            "title": "Language",
            "description": "Qwen3-TTS language name, or `auto`."
          },
          "sample_rate": {
            "type": "integer",
            "title": "Sample Rate",
            "description": "Always 24000."
          },
          "duration_s": {
            "type": "number",
            "title": "Duration S",
            "description": "Length of the speech in seconds."
          },
          "canceled": {
            "type": "boolean",
            "title": "Canceled",
            "description": "Whether generation stopped early."
          },
          "timings": {
            "$ref": "#/components/schemas/_Timings"
          }
        },
        "type": "object",
        "required": [
          "file_id",
          "audio_bytes",
          "model",
          "voice",
          "language",
          "sample_rate",
          "duration_s",
          "canceled",
          "timings"
        ],
        "title": "SpeechResponse",
        "description": "Capability-specific ``result`` for a succeeded speech job (`wav`)."
      },
      "ValidationError": {
        "properties": {
          "loc": {
            "items": {
              "anyOf": [
                {
                  "type": "string"
                },
                {
                  "type": "integer"
                }
              ]
            },
            "type": "array",
            "title": "Location"
          },
          "msg": {
            "type": "string",
            "title": "Message"
          },
          "type": {
            "type": "string",
            "title": "Error Type"
          },
          "input": {
            "title": "Input"
          },
          "ctx": {
            "type": "object",
            "title": "Context"
          }
        },
        "type": "object",
        "required": [
          "loc",
          "msg",
          "type"
        ],
        "title": "ValidationError"
      },
      "Voice": {
        "properties": {
          "id": {
            "type": "string",
            "title": "Id",
            "description": "Pass as `voice` to /v1/audio/speech."
          },
          "object": {
            "type": "string",
            "const": "voice",
            "title": "Object",
            "default": "voice"
          },
          "name": {
            "anyOf": [
              {
                "type": "string"
              },
              {
                "type": "null"
              }
            ],
            "title": "Name"
          },
          "transcript": {
            "anyOf": [
              {
                "type": "string"
              },
              {
                "type": "null"
              }
            ],
            "title": "Transcript",
            "description": "The reference transcript (given, or transcribed by kiapi), or null when cloned from the embedding only."
          },
          "duration_s": {
            "type": "number",
            "title": "Duration",
            "description": "Reference length in seconds."
          },
          "created_at": {
            "type": "integer",
            "title": "Created at",
            "description": "Unix seconds."
          }
        },
        "type": "object",
        "required": [
          "id",
          "duration_s",
          "created_at"
        ],
        "title": "Voice",
        "description": "A voice registered for `/v1/audio/speech`."
      },
      "VoiceCreateRequest": {
        "properties": {
          "audio": {
            "$ref": "#/components/schemas/FileRef",
            "title": "Audio",
            "description": "Reference audio as a file reference: `{type: file_id, file_id}`, `{type: url, url}`, or `{type: data_url, data_url}`. WAV, FLAC, MP3, or OGG of one clear speaker; 5 to 15 seconds works well. It is converted to 24 kHz mono and cut to `KIAPI_SPEECH_MAX_REFERENCE_SECONDS` (default 30)."
          },
          "name": {
            "anyOf": [
              {
                "type": "string",
                "maxLength": 200
              },
              {
                "type": "null"
              }
            ],
            "title": "Name",
            "description": "Label shown in the voice list. Omit it to use the uploaded file's name without its extension, or the start of the transcript when there is no file name (a data URL)."
          },
          "transcript": {
            "anyOf": [
              {
                "type": "string",
                "maxLength": 2000
              },
              {
                "type": "null"
              }
            ],
            "title": "Transcript",
            "description": "What the reference says, word for word. Omit it to have kiapi transcribe the reference (`KIAPI_SPEECH_TRANSCRIBE_REFERENCE`); without a transcript the voice is cloned from the speaker embedding only and sounds less like the reference."
          }
        },
        "type": "object",
        "required": [
          "audio"
        ],
        "title": "VoiceCreateRequest",
        "description": "Reference audio (and its transcript) to clone a voice from."
      },
      "VoiceList": {
        "properties": {
          "object": {
            "type": "string",
            "const": "list",
            "title": "Object",
            "default": "list"
          },
          "data": {
            "items": {
              "$ref": "#/components/schemas/Voice"
            },
            "type": "array",
            "title": "Data"
          }
        },
        "type": "object",
        "required": [
          "data"
        ],
        "title": "VoiceList"
      },
      "VoiceUpdateRequest": {
        "properties": {
          "name": {
            "anyOf": [
              {
                "type": "string",
                "maxLength": 200
              },
              {
                "type": "null"
              }
            ],
            "title": "Name",
            "description": "New label. Empty or null clears it."
          },
          "transcript": {
            "anyOf": [
              {
                "type": "string",
                "maxLength": 2000
              },
              {
                "type": "null"
              }
            ],
            "title": "Transcript",
            "description": "What the reference says, word for word. Used by the next speech in this voice. Empty or null clones from the speaker embedding only."
          }
        },
        "type": "object",
        "title": "VoiceUpdateRequest",
        "description": "Fields to change on a registered voice; omitted fields stay as they are."
      },
      "_Timings": {
        "properties": {
          "first_chunk_s": {
            "anyOf": [
              {
                "type": "number"
              },
              {
                "type": "null"
              }
            ],
            "title": "First Chunk S",
            "description": "Seconds until the first audio chunk."
          },
          "total_s": {
            "type": "number",
            "title": "Total S",
            "description": "Wall-clock generation time in seconds."
          }
        },
        "type": "object",
        "required": [
          "first_chunk_s",
          "total_s"
        ],
        "title": "_Timings"
      }
    }
  },
  "x-kiapi-capability": "speech",
  "x-kiapi-domain": "audio",
  "x-kiapi-root-openapi": "/openapi.json"
}