Lucidya Ltd Audio Transcription API

openapi: 3.2.0
info:
  version: 0.1.0
  title: Lucidya Public AI Audio Transcription API
  summary: This is the Public API Reference for the Lucidya **AI** Product.
  description: Our AI API enables you to get insightful analytics and predictions from your text batches.
  termsOfService: https://lucidya.com/service-agreement/
  contact:
    name: Lucidya Support
    url: https://lucidya.com/
    email: customer.support@lucidya.com
servers:
- url: https://api.lucidya.com
  description: Production
tags:
- name: Audio Transcription
paths:
  /audio_transcription/transcribe_offline:
    post:
      tags:
      - Audio Transcription
      summary: Submit Audio for Analysis
      description: 'Upload an audio file for transcription and analysis. The service processes the audio asynchronously — use the returned `job_id` to poll for status and retrieve results.


        **Supported audio formats:** MP3, MP4, MP4A, WAV, AIFF, FLAC, M4A, ALAC, WavPack, AAC, Ogg Vorbis, Opus, PCM.


        **Rate limit:** The total duration of submitted audio files must not exceed **100 minutes per minute**. If you exceed this limit, subsequent requests will be denied until the start of the next minute. This applies whether you submit one long file or multiple shorter files.


        **Next steps after submission:**

        1. [Check Audio Analysis Status](../audio_transcription/check_status) — lightweight status poll (`completed` or `in_progress`).

        2. [Get Audio Analysis Result](../audio_transcription/get_transcription_result) — full transcription with dialects, sentiments, and themes once status is `completed`.'
      operationId: post_audio_transcription_transcribe_offline
      parameters:
      - name: luc-authorization
        in: header
        description: API authorization token.
        required: true
        schema:
          type: string
      - name: num_speakers
        in: query
        description: The number of speakers in the audio file. Used for speaker diarization.
        required: true
        schema:
          type: integer
          minimum: 1
      - name: language
        in: query
        description: The language of the audio (e.g. "English", "Arabic"). If omitted, the language is detected automatically.
        required: false
        schema:
          type: string
      requestBody:
        required: true
        content:
          multipart/form-data:
            schema:
              type: object
              properties:
                file:
                  type: string
                  description: The audio file to transcribe.
                  contentMediaType: application/octet-stream
              required:
              - file
      responses:
        '200':
          description: Audio file accepted for processing.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AudioSubmissionResponse'
              examples:
                with_language:
                  summary: Language specified by the caller
                  value:
                    job_id: abc123
                    message: File submitted successfully. Transcription in progress.
                without_language:
                  summary: Language auto-detected
                  value:
                    job_id: abc123
                    message: File submitted successfully, transcription is in progress. Since no language was specified, the ai will detect the language automatically.
        '400':
          description: Bad Request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AudioErrorResponse'
              examples:
                invalid_format:
                  summary: Invalid audio format
                  value:
                    error: Invalid audio format.
                invalid_company:
                  summary: Incorrect company ID
                  value:
                    error: The company id is incorrect.
                duration_exceeded:
                  summary: File duration exceeds limit
                  value:
                    error: File duration exceeds the limit.
                missing_speakers:
                  summary: Number of speakers missing
                  value:
                    error: The Number of Speakers parameter is missing.
                unsupported_language:
                  summary: Unsupported language specified
                  value:
                    error: The specified language is not supported.
        '401':
          description: Unauthorized
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AudioErrorResponse'
              examples:
                unauthorized:
                  value:
                    error: Invalid or missing authorization token.
        '429':
          description: Too Many Requests
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AudioErrorResponse'
              examples:
                audio_duration_limit:
                  summary: Audio duration rate limit
                  value:
                    error: You have exceeded the allowed limit of 100 minutes of audio transcription per minute. Please try again after a minute.
                request_count_limit:
                  summary: Request count rate limit
                  value:
                    error: You have exceeded the allowed limit of 100 requests per minute. Please try again after a minute.
      servers:
      - url: https://api.lucidya.com
        description: Production
  /audio_transcription/check_status:
    get:
      tags:
      - Audio Transcription
      summary: Check Audio Analysis Status
      description: 'Check the processing status of a submitted audio transcription job. Returns only the job ID and current status — no transcript or analysis data.


        Use this endpoint for lightweight polling. Once the status is `completed`, call [Get Audio Analysis Result](../audio_transcription/get_transcription_result) to retrieve the full transcript, dialects, sentiments, and themes.'
      operationId: checkAudioAnalysisStatus
      parameters:
      - name: luc-authorization
        in: header
        description: API authorization token.
        required: true
        schema:
          type: string
      - name: id
        in: query
        description: The job ID returned from the POST /audio_transcription/transcribe_offline submission.
        required: true
        schema:
          type: string
      responses:
        '200':
          description: Job completed.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AudioStatusResponse'
              examples:
                completed:
                  value:
                    job_id: d38ee71d-b72b-4e28-9962-430a1612bb93
                    status: completed
        '202':
          description: Job still in progress.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AudioStatusResponse'
              examples:
                in_progress:
                  value:
                    job_id: d38ee71d-b72b-4e28-9962-430a1612bb93
                    status: in_progress
        '404':
          description: Job not found.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AudioErrorResponse'
              examples:
                not_found:
                  value:
                    error: Job ID not found. It may have been removed if it is too old.
      servers:
      - url: https://api.lucidya.com
        description: Production
  /audio_transcription/get_transcription_result:
    get:
      tags:
      - Audio Transcription
      summary: Get Audio Analysis Result
      description: 'Retrieve the full transcription and analysis result for a completed audio job.


        Returns the transcript with speaker diarization, dialect detection, sentiment analysis, and theme classification. If the job is still processing, the response will be `202` with an empty result.


        Use [Check Audio Analysis Status](../audio_transcription/check_status) for lightweight polling before calling this endpoint.'
      operationId: getAudioAnalysisResult
      parameters:
      - name: luc-authorization
        in: header
        description: API authorization token.
        required: true
        schema:
          type: string
      - name: id
        in: query
        description: The job ID returned from the POST /audio_transcription/transcribe_offline submission.
        required: true
        schema:
          type: string
      responses:
        '200':
          description: Transcription completed. Full analysis result returned.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AudioAnalysisResult'
              examples:
                completed:
                  summary: Completed single-speaker analysis
                  value:
                    job_id: d38ee71d-b72b-4e28-9962-430a1612bb93
                    status: completed
                    result:
                      transcripts_json:
                      - text: طيب مرحبا
                        speaker: 0
                        begins: 1280
                        ends: 1972
                      - text: أهلا وسهلا
                        speaker: 1
                        begins: 3288
                        ends: 3946
                      dialects:
                        speaker0:
                          dialect: egyptian
                          sub_dialect: egyptian
                        speaker1:
                          dialect: iraqi
                          sub_dialect: iraqi
                      sentiments:
                        speaker0: positive
                        speaker1: neutral
                      themes:
                        speaker0:
                          themes: complaint
                          sub_themes: awad
                        speaker1:
                          themes: question
                          sub_themes: Gamingggm
        '202':
          description: Transcription still in progress.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AudioAnalysisInProgress'
              examples:
                in_progress:
                  value:
                    job_id: d38ee71d-b72b-4e28-9962-430a1612bb93
                    status: in_progress
                    result: ''
        '400':
          description: Missing job ID parameter.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AudioErrorResponse'
              examples:
                missing_id:
                  value:
                    error: Missing job ID parameter.
        '404':
          description: Job not found.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AudioErrorResponse'
              examples:
                not_found:
                  value:
                    error: Job ID not found. It may have been removed if it is too old.
      servers:
      - url: https://api.lucidya.com
        description: Production
components:
  schemas:
    SpeakerThemes:
      type: object
      title: SpeakerThemes
      description: Theme classification result for a single speaker.
      properties:
        themes:
          type: string
          description: Primary theme classification (e.g. "complaint", "question").
        sub_themes:
          type: string
          description: Sub-theme classification.
      required:
      - themes
      - sub_themes
    AudioSubmissionResponse:
      type: object
      title: AudioSubmissionResponse
      description: Response returned when an audio file is successfully submitted for processing.
      properties:
        job_id:
          type: string
          description: Unique identifier for the transcription job. Use this to poll status and retrieve results.
        message:
          type: string
          description: Human-readable confirmation message.
      required:
      - job_id
      - message
    AudioErrorResponse:
      type: object
      title: AudioErrorResponse
      description: Error response returned by audio transcription endpoints.
      properties:
        error:
          type: string
          description: Human-readable error message.
      required:
      - error
    TranscriptSegment:
      type: object
      title: TranscriptSegment
      description: A single segment of transcribed speech.
      properties:
        text:
          type: string
          description: Transcribed text for this segment.
        speaker:
          type: integer
          description: Zero-indexed speaker identifier.
        begins:
          type: integer
          description: Segment start time in milliseconds from the beginning of the audio.
        ends:
          type: integer
          description: Segment end time in milliseconds from the beginning of the audio.
      required:
      - text
      - speaker
      - begins
      - ends
    AudioAnalysisInProgress:
      type: object
      title: AudioAnalysisInProgress
      description: Returned when the transcription job is still being processed.
      properties:
        job_id:
          type: string
        status:
          const: in_progress
        result:
          type: string
          description: Empty string while processing is in progress.
      required:
      - job_id
      - status
      - result
    AudioStatusResponse:
      type: object
      title: AudioStatusResponse
      description: Lightweight status-only response for a transcription job.
      properties:
        job_id:
          type: string
        status:
          type: string
          enum:
          - completed
          - in_progress
      required:
      - job_id
      - status
    SpeakerDialect:
      type: object
      title: SpeakerDialect
      description: Dialect detection result for a single speaker.
      properties:
        dialect:
          type: string
          description: Primary dialect classification (e.g. "egyptian", "iraqi", "saudi").
        sub_dialect:
          type: string
          description: More specific sub-dialect classification (e.g. "najdi", "egyptian").
      required:
      - dialect
      - sub_dialect
    AudioAnalysisResult:
      type: object
      title: AudioAnalysisResult
      description: Full transcription and analysis result for a completed audio job.
      properties:
        job_id:
          type: string
        status:
          const: completed
        result:
          type: object
          properties:
            transcripts_json:
              type: array
              description: Ordered list of transcript segments with speaker labels and timestamps (in milliseconds).
              items:
                $ref: '#/components/schemas/TranscriptSegment'
            dialects:
              type: object
              description: Detected Arabic dialect per speaker. Keys are `speaker0`, `speaker1`, etc.
              additionalProperties:
                $ref: '#/components/schemas/SpeakerDialect'
            sentiments:
              type: object
              description: Overall sentiment per speaker. Keys are `speaker0`, `speaker1`, etc.
              additionalProperties:
                type: string
                enum:
                - positive
                - neutral
                - negative
            themes:
              type: object
              description: Detected themes per speaker. Keys are `speaker0`, `speaker1`, etc.
              additionalProperties:
                $ref: '#/components/schemas/SpeakerThemes'
          required:
          - transcripts_json
          - dialects
          - sentiments
          - themes
      required:
      - job_id
      - status
      - result