Transcription guide

Getting started with Transcription

Transcription turns speech into text with Whisper, on your own device. Nothing is uploaded.

Open Transcription

Get the speech model once

The first time, the page says This tool needs a model. Choose Get more…: the dialog shows Whisper tiny, its licence and its size, about 40 MB. Choose Download. It is stored on this device, so you only do this once.

Add audio and transcribe it

1. Drop audio or video files on the page, or choose Choose files. MP3, WAV, M4A, OGG, FLAC and video files work, several at once. Or choose Record to use the microphone.
2. Leave Spoken language on Detect, or pick the language.
3. Choose Transcribe. Each file is read in turn and marked Done.

Tip: Whisper tiny is quick but can get names and noisy audio wrong. Read the text before you use it.

Correct and save the text

Pick a file in the Files list to see its text. Type in the box to correct it; your corrections are kept for Text and Word. Show times adds the time of each line.

Choose a format: Text (.txt), Word (.docx), Subtitles (.srt) or Web subtitles (.vtt), then choose Download. Copy, Save to Files and Open in Documents are there too.