Captioner

Captioner leans on a fine-tuned Whisper model to transcribe video audio rather than skim it. The team claims 98%+ accuracy and tunes the engine hardest for English and Chinese, with added support for Cantonese. Accuracy does vary by language, so testing a short clip first is worth the few minutes.

Not every transcription tool hands back timing you can trust. Captioner adds a processing pass that lines subtitles up with the spoken audio, then offers word-level timestamps when you want finer control. That saves the tedious step of nudging caption blocks by hand.

After transcription, everything happens in the browser. You fix wording, adjust timing, and reflow captions without opening another app. The editing tools were built by creators, so the layout is aimed at the parts people actually change, not at a wall of settings.

If you already wrote a script or transcript, you can import it and let Captioner match it to the audio. It works like YouTube's auto-sync feature but keeps the text editable, which helps when your version is more accurate than the machine's.

One click turns subtitles into another language, and it works as a subtitle translator for clips you plan to republish. Captioner covers target languages such as Spanish, Portuguese, French, German, Russian, Arabic, Hindi, Thai, and more, so the same clip can serve several audiences without a second editing pass.

Captioner exports SBV, SRT, and VTT files, the formats most editors read. If you'd rather ship the video with captions baked in, pick a font style in the editor and render an open-caption video directly, so viewers see text even with captions switched off.

A built-in voice extraction step separates speech from music and noise before transcription. You don't have to mute the soundtrack first, and captions stay readable on videos that carry background audio.