Everything Captionshaper does

From the first word it hears to the finished video, all on your own computer.

It listens properly

Speech recognition runs on your Mac or PC with OpenAI's Whisper model, so it understands accents, fast talkers and around 99 languages.

  • Finds the speech firstVoice detection skips intros, music and silence, so no words get invented over a beat drop.
  • Timed to the wordEach word starts when the voice does. Captions appear in sync, and word highlights land on the beat.
  • Detects the languageOr pick it yourself. It can also turn any supported language into English captions.
  • Three modelsRecommended for most videos, Most accurate for harder languages and audio, Fast for quick drafts.
It listens properly

Captions that read well

A transcript isn't captions. Captionshaper decides where every caption starts and ends, and where each line breaks.

  • Natural line breaksBreaks fall at sentence and clause ends, never after “the”, “a” or “to”.
  • Never across a pauseWhen the speaker stops, the caption stops. Lines fit the width of your video.
  • Easy to fixWords it isn't sure about are underlined. Edit in place, find and replace, split, merge and undo.
  • Saved as you goClose the app and open the same video later: your edits are still there.
Captions that read well

Styles made for social video

Nine looks to start from and eleven caption fonts. The preview is exactly what gets exported.

  • Highlight the spoken wordChange its colour or put it on a box, the way the big creators do.
  • Word animationsPop, fade in or slide up. Hide or dim the words that haven't been said yet.
  • Outline, shadow and glowEach with its own colour, plus a background box behind lines.
  • Phone viewSee where TikTok, Reels and Shorts put their buttons, so captions never hide under them.
Styles made for social video

Every way out

Take your captions wherever they need to go.

  • Video with captionsAn MP4 with the captions burned in, ready to post.
  • Subtitle filesSRT for YouTube and video editors, VTT for the web, ASS with your full style.
  • TranscriptPlain text, split into paragraphs at long pauses.
  • Private by designNothing is uploaded. After the one-time model download, it works offline.
Every way out

Captions that keep up with you.

macOS 12+ on Apple Silicon · Windows 10 and 11 (64-bit)