name: Deepgram Speech-to-Text API Capabilities
description: >-
  Capabilities exposed by Deepgram's Speech-to-Text API for batch and
  streaming transcription with the Nova model family.
capabilities:
  - name: Transcribe Pre-Recorded Audio
    description: Submit hosted or uploaded audio for transcription.
    operationId: transcribePreRecorded
    method: POST
    path: /v1/listen
  - name: Transcribe Streaming Audio
    description: Open a WebSocket session for real-time transcription.
    operationId: transcribeStreaming
    method: WEBSOCKET
    path: /v1/listen
  - name: Detect Languages
    description: Detect the spoken language in an audio source.
    operationId: detectLanguage
    method: POST
    path: /v1/listen
  - name: Diarize Speakers
    description: Tag transcript segments with speaker identifiers.
    operationId: diarize
    method: POST
    path: /v1/listen
  - name: Smart Format Transcripts
    description: Apply punctuation and smart formatting to transcripts.
    operationId: smartFormat
    method: POST
    path: /v1/listen
use_cases:
  - name: Call center transcription and analytics
    description: Transcribe customer calls and feed text to QA, sentiment, and topic models.
  - name: Meeting and interview transcription
    description: Generate searchable transcripts from recorded meetings, interviews, and podcasts.
  - name: Media captioning
    description: Produce captions and subtitles for video and live broadcasts.
  - name: Voice agent ASR
    description: Provide low-latency streaming transcription for conversational AI.