Silo Caption.
Private video captions for Mac.
Silo Caption works on video files you already have. It does not record your screen or microphone and it does not write meeting notes. To record and transcribe a lecture or interview, use Silo Transcribe; for meeting notes, use Silo Meet.
Captions from footage you already have.
Import an MP4 or MOV from the sidebar, with File > Import Video (Command-I), or by opening it with Silo Caption from Finder. A speech model running on your Mac turns the dialogue into timed cues you can correct. Choose the spoken language in Settings or leave it on auto-detect. The model downloads once, after you approve it.
Karaoke highlights, word by word.
Set any font installed on your Mac, its size, text, background and border colors with opacity, border width, and left, center or right alignment. Karaoke animation lights each word on the timing it was recognized with. Captions you edit or translate, and cues without matching word timing, stay static instead of guessing.
Translate the saved revision.
Pick a source and target language your Mac supports and translate saved captions with Apple's on-device translation. Approve a language download if macOS asks for one. The Output menu switches between the original captions and each saved translation for preview and export.
Save a caption project and pick it up again from Saved projects. Preview plays the saved captions over your video in the shape and resolution chosen for export. Unsaved edits hold preview and export until you save.
Render a new MP4 or MOV in landscape, portrait or square, at Standard (1080-pixel short edge) or High (2160-pixel short edge), or save SRT and VTT files. Audio is re-encoded as AAC, an existing file is never replaced, and Cancel stays available while a video renders.
Video, audio and caption text are processed locally, with no Silo account. Silo Caption asks for no microphone, camera or Screen Recording access. The speech model and any Apple language downloads need an internet connection. Apple may receive translation usage and performance metrics, subject to your system sharing choices, but not the original or translated text.