Break on sentence flow
Cues follow the shape of speech and never split a word. Oversized passages break where a viewer naturally pauses.
Turn video and audio into properly timed subtitles. The speech model is bundled inside the app—nothing to download, and nothing ever leaves your Mac.
TRANSCRIPTION IS ONLY THE START
Raw speech recognition produces long blocks with arbitrary timing. Real subtitles have rules: they must read naturally, fit the frame, remain visible long enough, and hand off cleanly to the next cue.
SubtitleForge applies those rules after transcription, turning words into cues made for watching.
Cues follow the shape of speech and never split a word. Oversized passages break where a viewer naturally pauses.
Line breaks are redistributed to avoid one long line floating above a short one. The picture stays calm and readable.
When a cue is too fast, SubtitleForge extends its timing instead of cutting words. Reading speed is measured in characters per second.
No cue remains on screen when the next begins. Every cue also respects your minimum and maximum duration.
REVIEW WITH EVIDENCE
The review pass lists cues that read too fast, sit too briefly, run too long, overflow a line, or overlap the next. Every warning includes the exact figure, so you can fix with confidence instead of scrubbing and guessing.
YOUR RULES, BUILT IN
Start with dependable broadcast settings: 42 characters per line, two lines, and 17 characters per second. Then adjust line length and reading speed for your delivery format.
Move every cue earlier or later with one timing control. Useful when the edit changes but the words do not.
Deliver SubRip .srt, WebVTT .vtt, or a clean plain transcript.
ONE APP. ONE PURCHASE.
SubtitleForge runs entirely on your Mac, with the speech model included. No account. No cloud queue. No subscription.