Edit transcripts

Click any transcript in your library to open it. The transcript is on the left, with the model and language menus above it and the player below. The right panel has three tabs: Overview, Chat and Summary. Hide or show the panel with the sidebar button at the top right of the window.

The right panel

• Overview: a video or audio preview at the top, and below it four tabs:
  • Settings: name, tags and how the transcript looks (see Appearance options).
  • Speakers: run speaker identification and rename speakers. See speaker diarization.
  • Translate: translate the transcript (Pro). See languages and translation.
  • Info: title, description, duration, word count, model, language and dates.
• Chat: ask an AI questions about this transcript (Pro). See chat.
• Summary: a Summary tab where you click Generate to create an AI summary (Pro) and then edit it, and a Notes tab for your own notes. See summary prompt.

Text and Segments views

Switch views with Display Mode in Overview › Settings.

• Segments shows the transcript as paragraphs with speaker names and timestamps. This is the view you edit in.
• Text shows one continuous block of text for reading. It is read-only.

Edit the text

1. Make sure Display Mode is Segments and Allow Edit is on in Overview › Settings.
2. Click into a paragraph and type.
3. Press Return, click outside the paragraph, or click the checkmark (Save changes) in the paragraph’s action bar to save. Press Esc to discard your changes.

Hover over a paragraph to see its action bar:

• Play from here (or Pause while it’s playing).
• Assign speaker: pick a speaker for the paragraph, or Remove speaker.
• Favorite: star the paragraph.
• Copy paragraph.
• Select: start selecting paragraphs to assign a speaker to several at once (Assign Speaker) or delete them (Delete All). Click Exit when you’re done.
• More actions: New Segment, Merge Up and Merge Down.
• Delete paragraph.

Note: To add or merge segments, turn off Group by Speaker first, so each row is a single segment. Deleting segments can’t be undone.

Edit timestamps

With Show Timestamps on, each paragraph shows its start and end time. Click the time (Edit timestamp) to open a small editor and adjust the start and end. A paragraph can’t start before the previous one ends or end after the next one starts.

Playback

The player at the bottom of the transcript has play and pause, previous and next segment buttons, a speed menu (0.5x, 1.0x, 1.5x, 2.0x) and mute. For videos, you can show or hide subtitles on the preview.

• Space plays and pauses.
• ⌘ ← and ⌘ → jump to the previous or next segment.

Three settings in Overview › Settings › Playback control how the transcript follows the audio:

• Click to Play: click a word to play from that point.
• Auto-scroll: keep the transcript scrolled to the current segment during playback.
• Highlight: Word, Segment or None. Word appears only when the model produced word timestamps.

Favorites

Star important paragraphs with Favorite in the action bar. Turn on Favorites Only in Overview › Settings to show only starred paragraphs, and use Show Favorites Only when you export to export just those.

Search and replace

1. Type in the Search... field in the toolbar. Matches are highlighted, with a count such as 2 / 7.
2. Use Previous and Next to move between matches, or press Return.
3. To replace, click the swap button (Toggle Replace) inside the search field, type the new text in Replace..., then click Replace All.

Replace All ignores capitalization and also changes matching text in the translation. To fix the same mistake in future transcripts automatically, add a text replacement.

Reflow transcripts

Reflow rebuilds the segments so each line has a set maximum length. It’s useful before exporting SRT or VTT subtitles.

1. Click the More options menu (…) next to Restart and choose Reflow Transcripts….
2. Choose Maximum CPL (characters per line, 20 to 120, 40 to 80 recommended) or Maximum WPL (words per line, 3 to 20, 8 to 12 recommended).
3. Set the value with the slider and click Reflow.

Note: Reflowing overwrites all edits you made to the segments. It needs word-level timestamps, so the menu item is disabled for transcripts without them, such as Apple Intelligence transcripts or Mistral transcripts with diarization on.

Generate an AI title and description

Choose Generate AI title and description from the More options menu, or right-click the transcript in the library. TranscribeX uses your default AI service to write a title and a short description, shown in the Info tab. To do this automatically after every transcription, turn on Settings › Transcription › Use AI to Auto-generate Title and Description. See AI services.

To rename a transcript yourself, edit the name in Overview › Settings.

More options menu

The More options menu (…) next to Start or Restart also has:

• Show source file in Finder and Delete source file.
• Open web link, for YouTube and other downloads.
• Copy Transcripts: copies the whole transcript, with translations if they’re shown.
• Delete: choose Delete Job to remove only the transcript, or Delete Job and Audio/Video File… to also remove the media file (moved to the Trash if Move to Trash is on in Settings › General).

Click Restart to transcribe the file again, for example with a different model or language. This replaces the current transcript and your edits.

Appearance options

These are saved per transcript in Overview › Settings:

The segment options appear only in Segments view. If a model doesn’t produce timestamps, timestamps and speakers are turned off for that transcript.