Speaker names in a transcript, exporting and sharing

Naming and merging speakers

• Naming: Click a label in the text and type. It changes in all of that person’s paragraphs. Left empty, it goes back to the number.
• One person under two numbers: Give both labels the same name and they become one person. Their voice prints merge too, so later speech is recognised as one. The method errs on the side of splitting, so this is common.
• Two people under one number: Naming cannot separate them. There is also no button that moves a paragraph to another speaker, so mark the parts that matter by hand in the text.

Names are kept with the transcript, and people you share it with see the same ones. AI transcription accuracy explains why separation behaves this way.

Export formats compared

The Export menu builds the file in your browser and downloads it. The download icon beside each transcript in the list does the same.

The speaker’s name is written only when the speaker changes. The subtitle generator covers subtitle files, and audio to Word transcription covers documents. Transcripts are deleted when your plan’s retention period ends, so download what you want to keep.

Sharing with your team

If you are in a team, use Share at the top of the transcript page to choose which teams to share with. Members open the text and edit at the same time, seeing each other’s cursors and names. You can work together while transcription runs as well.

Only the owner changes sharing or deletes a transcript. When the list grows, the Select button lets you share several at once. Opening a team needs a team plan: See pricing.

The public link

In the same window, turning on Public link opens the transcript’s own address to people without an account. They can read and download it, and see the media if it comes from a stream, but they cannot change it. This right belongs to team plans and team members.

The page is closed to search engines, but anyone with the address can read it. Do not use it for text with personal data: See KVKK and voice recordings. Turn it off and the link stops working.

Version history: Going back

More → Versions keeps earlier states of the transcript. Choose one to see what would change: Green comes back, red goes. Restore this version first saves the current state as a version too, so going back loses nothing. The change reaches everyone with it open. You cannot restore while transcribing, so stop first. Old versions are removed after a period given in the privacy notice.

Original text: What the engine wrote first

More → Original text keeps what the engine wrote first. Show original marks in red what was in the original and changed, and in green what was written later. Find deleted lists removed passages of at least three words with their time and speaker, and clicking a time takes the recording there.

This helps to find a passage deleted by mistake, or to show what you changed in research. If you are anonymising, remember that old names still show here and in versions, so export a clean file to share it. For the first steps, go back to your first transcript.