Skip to content

What’s new in 2.1

The short version: your transcript is now saved with your project, separately from your markers. Close the panel, switch comps, clear every marker, hand the project to someone else — your captions are still there in the Navigate tab.

New to the panel? Open VoxMark and choose ? ▸ Show the tour.

  • Saved transcripts. Every transcription or SRT import is saved in a Chroma/VoxMark folder beside your project file. See Your project’s VoxMark data.
  • Each project keeps its own, even when several projects share a folder. Save a new version (Edit_v2 from Edit_v1) and it picks up the earlier version’s transcript.
  • Nothing is simply lost. Transcribing again, or importing an SRT over an existing transcript, first keeps the old one in a replaced folder.
  • Clearing markers keeps the transcript. Clear All and Save & Clear remove the markers but leave every caption in the Navigate tab, ready to recreate them.
  • Transcript only. Tick Transcript only (no markers) to transcribe without touching your timeline.
  • Create markers from transcript adds them whenever you want — from the Transcribe tab, or the ⊕ button in the Navigate tab. Run it again and it replaces its own markers; markers you placed yourself are left alone.
  • Large jobs (500+ captions) create their markers in batches with visible progress.

Double-click any caption in the Navigate tab to fix a word. Press Enter to keep it — the saved transcript and the marker on your timeline both update. See The Navigate tab.

Editing a caption in place: the transcription heard “Boxmark”, corrected here to “VoxMark”

Whisper tends to start captions early around pauses — sometimes by half a second, a dozen frames at 24 fps. VoxMark now corrects each caption against the word timings and, where those are wrong too, against the audio itself, so markers land on the words.

  • Choose your Groq model: Whisper Large v3 Turbo (the default — faster, more accurate timing) or Whisper Large v3 (slightly better word for word).
  • Groq’s paid developer tier now gets its full 100 MB — roughly 50 minutes of audio.
  • A provider that refuses a file as too large now says so, with the real limits.
  • SRT export and Convert to Captions land at the right time when your audio doesn’t start at the beginning of the timeline.
  • Premiere: hiding marker text no longer empties the Navigate tab when you switch sequences or reopen the panel.
  • After Effects 2026: Set work area to caption puts the work area on the caption, not at the start of the comp. In Premiere the same button is Set In/Out to caption.
  • Switching between projects always loads the right transcript.

The ? menu: the VoxMark guide, What’s new, and Show the tour

The new ? button in the panel header opens this guide and replays the tour. The tour starts on its own the first time you open VoxMark: it begins at Scan — solo your voiceover first — walks the Transcribe settings, lets you try the Navigate tab on a sample transcript — with real markers, in a temporary “VoxMark Tour” comp that is removed when the tour ends — editing a caption included, and finishes with export and import.

A step of the tour, explaining that only AssemblyAI tells voices apart