Whisper API

AI 모델

OpenAI Whisper API를 통한 빠른 클라우드 전사

할 수 있는 것

  • Cloud transcription — Fast speech-to-text via OpenAI's API (often faster than local for large files)
  • Language specification — Set expected language for better accuracy on non-English audio
  • Custom prompts — Provide speaker names, technical terms, or jargon for improved accuracy
  • Format options — JSON (with timestamps) or plain text output
  • Auto file output — Saves transcription alongside the original audio file

    • 시도해볼 질문

  • "Transcribe this interview.mp3 via the API"
  • "Transcribe this meeting in Spanish"
  • "Transcribe with context: speakers are Dr. Smith and Prof. Jones discussing quantum computing"
  • "Get JSON transcription with timestamps"
  • "Transcribe this earnings call with company-specific terminology hints"

    • 전문가 팁

  • Requires OPENAI_API_KEY environment variable
  • Custom prompts dramatically improve accuracy for domain-specific content
  • Language hints help when audio quality is poor or accented
  • API is typically faster than local processing for files over 10 minutes
  • Max file size is 25MB — use local Whisper for larger files
  • JSON format includes segment timestamps useful for time-coded summaries