No upload · No watermark · No minute quotas
Karaoke caption generator
Karaoke captions — where each word lights up as it's spoken — need word-level timestamps, not sentence blocks. CaptionForge gets them straight from Whisper running on your device, then animates them in four modes: highlight, sweep, pop and reveal. The karaoke preset is free, because shipping a caption tool without karaoke would be shipping it broken.
- True word-level timing from Whisper (±120ms on clean speech)
- 4 karaoke modes: highlight, sweep, pop, reveal
- Karaoke included in the free tier — not paywalled
- Word edits keep sync: timings redistribute when you fix text
How it works
- 1 Drop a video — speech is transcribed locally with per-word timestamps.
- 2 Pick a karaoke mode: highlight, sweep, pop or reveal.
- 3 Adjust colors and position; fix any misheard words.
- 4 Export the MP4 — animation timing matches the preview exactly.
Frequently asked questions
What are karaoke captions?
Captions where the spoken word is visually emphasized in real time — like a karaoke machine's bouncing ball. It's the dominant style in short-form video because viewers' eyes track the active word, which keeps attention on muted feeds.
What's the difference between the four modes?
Highlight changes the active word's color or background pill as it's spoken. Sweep moves a continuous color fill across the line, karaoke-machine style. Pop scales each word up as it lands. Reveal shows words one by one as they're spoken. All four use the same word-level timing; they differ in how the emphasis is drawn.
Why does word-level timing matter?
Sentence-level subtitle files (like a plain SRT) only know when a line starts and ends — animating words from that means guessing, and the highlight drifts off the actual speech. Whisper outputs a timestamp for every word, so the animation locks to the audio. That's also why SRT imports offer a one-click local re-transcribe to unlock karaoke.
Is karaoke really free? Most tools gate animated styles.
The karaoke highlight preset is free, along with three other presets — word-by-word animation is the expected format, and gating it would make the free tier feel broken. Studio ($9/mo) adds the other presets, all four karaoke modes with a custom style editor, saved presets and brand fonts.
What happens to the timing when I edit a word?
Fixing text redistributes the word timings within the segment proportionally, so the karaoke stays in sync — you don't re-time anything by hand. Boundary nudging is there if you want manual control.