HOW TO USE THIS TOOL
Turn a recording into a transcript
Choose one M4A (AAC), MP3 or PCM WAV recording up to 120 minutes and 512MB, select the spoken language and start. The first run downloads about 95MB of model and program files from this site. Cached files can be reused; there is no subscription or per-use credit.
Audio is decoded in small sections and processed on your PC with Whisper-base. Your audio, filename and transcript are not uploaded. Each timestamp marks the start of an approximately 30-second section. Recognition can be wrong, especially with noise, overlapping speakers and words at section boundaries. Speaker identification and content summaries are not provided.
Pause or cancel keeps completed text. A temporary draft and processing position are saved in this browser, without the audio. To resume, choose the same recording. A new recording replaces the prior draft after confirmation. Clearing browser data can remove it; save important results.
Review and edit the text, copy it, or download TXT and editable DOCX. HWP and HWPX export is not provided. DOCX can be opened in Hancom Office, with compatibility depending on version. Keep the tab open during conversion; speed depends on your PC. This is a desktop beta; actual Mac and phone testing is pending.