🎵 Audio Transcription & Translation

Upload an audio file or use your microphone to transcribe or translate speech.

Model Selection
Task
Language (for transcription)
1 10
Examples
Audio Input Model Selection Task Language (for transcription) Include Timestamps Beam Size (Higher = Better Accuracy but Slower)

Features

  • Model Selection: Choose from 6 different Whisper models with speed/accuracy tradeoffs
  • Task Options: Transcribe audio in original language or translate to English
  • Language Selection: Auto-detect or specify input language for better accuracy
  • Multiple Input Methods: Upload audio files or record with microphone
  • Timestamps: Option to include word-level timestamps
  • Beam Search: Adjustable beam size for better accuracy

Model Information

Model Parameters Speed Best For
Whisper Tiny 39M Fastest Quick transcriptions, low resources
Whisper Base 74M Fast Balanced performance
Whisper Small 244M Medium Better accuracy
Whisper Medium 769M Slow High accuracy transcriptions
Whisper Large 1.5B Slower Very high accuracy
Whisper Large-v2 1.5B Slower Latest improvements
  • Supported Formats: WAV, MP3, M4A, FLAC
  • Note: First transcription may take 10-60 seconds (model loading)