Camb.AI Text-to-Speech API

openapi: 3.2.0
info:
  title: Camb.AI Dubbing Text To Speech API
  description: 'The Camb.AI API provides generative voice AI - text-to-speech, dubbing, translation, voice discovery and cloning, and transcription - powered by the MARS (speech / voice) and BOLI (translation) research models. The REST API is served from https://client.camb.ai/apis and authenticated with an x-api-key header. Most operations are asynchronous: a POST starts a task and returns a task_id, the task is polled with a GET until its status is SUCCESS, and the result (audio, transcript, or translation) is then retrieved. Real-time streaming for TTS, transcription, and speech-to-speech translation is offered over separate public WebSocket channels (see the companion AsyncAPI document). Endpoint paths are grounded in the public documentation at https://docs.camb.ai; request and response schemas below are modeled from the documented parameters and may be simplified.'
  version: '1.0'
  contact:
    name: Camb.AI
    url: https://www.camb.ai
servers:
- url: https://client.camb.ai/apis
  description: Camb.AI production API
security:
- apiKeyAuth: []
tags:
- name: Text-to-Speech
  description: Convert text to speech with the MARS voice models.
paths:
  /tts-stream:
    post:
      operationId: createTtsStream
      tags:
      - Text-to-Speech
      summary: Stream text-to-speech
      description: Synthesize speech from text and return a binary audio stream in real time. The response is an audio byte stream (for example audio/wav or audio/mpeg) and includes an X-Credits-Required header indicating credit usage.
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/TtsStreamRequest'
      responses:
        '200':
          description: A binary audio stream.
          content:
            audio/wav:
              schema:
                type: string
                format: binary
        '401':
          $ref: '#/components/responses/Unauthorized'
  /tts:
    post:
      operationId: createTts
      deprecated: true
      tags:
      - Text-to-Speech
      summary: Create a text-to-speech task (deprecated)
      description: Deprecated. For new integrations use POST /tts-stream. Starts an asynchronous text-to-speech task and returns a task_id to poll.
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/TtsRequest'
      responses:
        '200':
          description: Task accepted.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/TaskCreated'
        '401':
          $ref: '#/components/responses/Unauthorized'
  /tts/{task_id}:
    get:
      operationId: getTtsStatus
      tags:
      - Text-to-Speech
      summary: Poll a text-to-speech task
      description: Returns the status of a TTS task and, when complete, the run_id used to fetch the audio.
      parameters:
      - $ref: '#/components/parameters/TaskId'
      responses:
        '200':
          description: Current task status.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/TaskStatus'
        '401':
          $ref: '#/components/responses/Unauthorized'
  /tts-result/{run_id}:
    get:
      operationId: getTtsResult
      tags:
      - Text-to-Speech
      summary: Download text-to-speech audio
      description: Retrieves the generated audio file for a completed TTS run.
      parameters:
      - name: run_id
        in: path
        required: true
        description: The run identifier returned by GET /tts/{task_id}.
        schema:
          type: integer
      responses:
        '200':
          description: The generated audio file.
          content:
            audio/wav:
              schema:
                type: string
                format: binary
        '401':
          $ref: '#/components/responses/Unauthorized'
components:
  parameters:
    TaskId:
      name: task_id
      in: path
      required: true
      description: The task identifier returned when the task was created.
      schema:
        type: string
  responses:
    Unauthorized:
      description: Missing or invalid API key.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
  schemas:
    TaskCreated:
      type: object
      properties:
        task_id:
          type: string
          description: Identifier used to poll the task.
    Error:
      type: object
      properties:
        message:
          type: string
        code:
          type: string
    TtsRequest:
      type: object
      required:
      - text
      - voice_id
      - language
      properties:
        text:
          type: string
          description: The text to be converted to speech.
        voice_id:
          type: integer
          description: Voice identifier.
        language:
          type: string
          description: Locale tag such as en-us.
        gender:
          type: integer
          description: Speaker gender hint (0, 1, 2, or 9).
        age:
          type: integer
          description: Speaker age characteristic.
    TtsStreamRequest:
      type: object
      required:
      - text
      - language
      - voice_id
      properties:
        text:
          type: string
          minLength: 3
          maxLength: 3000
          description: The text to synthesize into speech.
        language:
          type: string
          description: BCP-47 locale code, for example en-us or fr-fr.
        voice_id:
          type: integer
          description: A voice profile ID from GET /list-voices.
        speech_model:
          type: string
          description: Model variant, for example mars-8.1-flash-beta or mars-instruct.
        user_instructions:
          type: string
          description: Guidance for overall delivery (mars-instruct only).
        enhance_named_entities_pronunciation:
          type: boolean
        output_configuration:
          type: object
          description: Output format, sample rate, and enhancement settings.
        voice_settings:
          type: object
          description: Speaking rate, accent preservation, and reference quality.
        inference_options:
          type: object
          description: Stability, temperature, and speaker similarity.
    TaskStatus:
      type: object
      properties:
        status:
          type: string
          description: Task status, for example PENDING, PROCESSING, SUCCESS, or FAILED.
          enum:
          - PENDING
          - PROCESSING
          - SUCCESS
          - FAILED
        run_id:
          type: integer
          description: Present when the task has completed; used to fetch the result.
  securitySchemes:
    apiKeyAuth:
      type: apiKey
      in: header
      name: x-api-key