openapi: "3.0.0"
info:
  title: 'Mimic 3'
  version: '0.1'
  description: 'A fast and local neural text to speech system for Mycroft'
schemes:
  - http
servers:
  - url: http://localhost:59125
    description: Local server
paths:
  /api/tts:
    get:
      summary: 'Speak text to WAV'
      parameters:
        - in: query
          name: text
          required: true
          description: 'Text to speak'
          schema:
            type: string
            example: 'Welcome to the world of speech synthesis!'
        - in: query
          name: voice
          description: 'Voice in the form <lang>/<name>_<quality> optional with #<speaker> at the end'
          schema:
            type: string
            example: 'en_UK/apope_low'
        - in: query
          name: noiseScale
          description: 'Volatility of speaker (0-1, default: 0.667)'
          schema:
            type: number
            example: 0.667
        - in: query
          name: noiseW
          description: 'Volatility of individual phonemes (0-1, default: 0.667)'
          schema:
            type: number
            example: 0.8
        - in: query
          name: lengthScale
          description: 'Speed of speaker (default: 1.0, faster &lt; 1 &lt; slower)'
          schema:
            type: number
            example: 1.0
        - in: query
          name: ssml
          description: 'Input text is SSML'
          schema:
            type: boolean
            example: false
      produces:
        - audio/wav
      responses:
        '200':
          description: audio
          schema:
            type: binary
    post:
      summary: 'Speak text to WAV'
      requestBody:
        required: true
        description: 'Text to speak'
        content:
          text/plain:
            schema:
              type: string
              example: 'Welcome to the world of speech synthesis!'
      parameters:
        - in: query
          name: voice
          description: 'Voice in the form <lang>/<name>_<quality> optional with #<speaker> at the end'
          schema:
            type: string
            example: 'en_UK/apope_low'
        - in: query
          name: noiseScale
          description: 'Volatility of speaker (0-1, default: 0.667)'
          schema:
            type: number
            example: 0.667
        - in: query
          name: noiseW
          description: 'Volatility of individual phonemes (0-1, default: 0.667)'
          schema:
            type: number
            example: 0.8
        - in: query
          name: lengthScale
          description: 'Speed of speaker (default: 1.0, faster &lt; 1 &lt; slower)'
          schema:
            type: number
            example: 1.0
        - in: query
          name: ssml
          description: 'Input text is SSML'
          schema:
            type: boolean
            example: false
      produces:
        - audio/wav
      responses:
        '200':
          description: audio
          schema:
            type: binary
  /api/voices:
    get:
      summary: 'Get available voices'
      produces:
        - application/json
      responses:
        '200':
          description: voices
          schema:
            type: object