Audio Transcript

Upload audio or voice messages and generate transcripts in English, Persian/Farsi, mixed speech, or auto-detect.

Transcribe audio

Supports MP3 from YouTube, WhatsApp and Telegram voice notes, interviews, lectures, and conversations. Transcription preserves the spoken language — it does not translate.

Long audio may take several minutes. Files over ~15 minutes are split into chunks automatically when ffmpeg is available.

Audio type

MP3, M4A, WAV, OGG, Opus, WebM, FLAC, and MP4 audio.

Audio file

Maximum upload size: 100 MB.

For long recordings, compress to MP3/M4A or split into shorter files.

Title (optional) Language

Auto-detect works well for English, Persian/Farsi, and mixed speech.

Speed / accuracy

Faster models run quicker; More accurate models are slower but better for noisy or mixed audio.

Clean up transcript

Improves punctuation and obvious speech-to-text mistakes. Raw transcript is always preserved.

Partial transcript

Optional. Transcribe only part of the audio.

Start at

hr min sec

End at

hr min sec