How to add captions to a video automatically

Attach the video to Recut and type "add captions". Recut listens to the speech, shows a few words at a time exactly as they are said, and burns the captions into the picture in a style that reads on a phone.

Step by step

1. Attach the videoOpen Recut (in the browser, or the Mac, Windows or Android app), start a project and drag the video in, or use the paperclip.
2. Ask for captionsType what you want in plain words. You can say the style (white text with a dark outline, or boxed on a dark box), the size (small, medium or large), where they sit (bottom, middle or top), the colour and how many words show at once.
3. Check the planRecut shows the step before anything runs. Drawing the captions happens on your device; timing them to the speech uses a transcription model, which costs a fraction of a cent.
4. Run it and comparePress Run, play the result, and ask for a change ("make them smaller", "move them to the top") if you need one. The earlier version stays on the timeline.

What Recut does

Recut transcribes the sound to get a time for every word, then groups the words into captions. A caption ends early at the end of a sentence, at a comma once it is half full, or before a pause of 0.6 seconds or more, so words never hang on screen over silence.

If you paste the script, Recut keeps your spelling (brand names, product names, slang) and still takes the timing from the audio. With no transcription available it can spread the words evenly across the video instead. The text is drawn on a canvas and laid over the video with ffmpeg, so captions look the same in every browser and app.

Tips

• Three to six words at a time reads well on a phone; four is a good start for TikTok and Reels.
• Bottom captions sit above the buttons and text that phone apps put on top of a video.
• If the video still needs resizing for a platform, do that first, so the captions are placed on the final frame.

What it doesn't do

• Captions are burned into the picture. Recut does not write a separate SRT or VTT subtitle file.
• A video gets at most 120 captions; a longer text is grouped into fewer, longer captions.
• There are no animated word-by-word highlight styles: the styles are outline and boxed.

Related guides

• How to transcribe a video to textAttach a video and ask for a transcript: Recut takes out the sound and runs a speech-to-text model such as ElevenLabs Scribe or Whisper.
• How to resize a video for TikTok, Reels and ShortsAsk Recut to export a video for TikTok, Instagram Reels or YouTube Shorts: a 9:16 frame at 1080×1920, 30 fps and even loudness.
• How to add an AI voiceover to a videoPaste a script, pick a voice, and Recut reads it with an AI voice model and lays the voiceover on your video, with music turned down under it.
• How to dub a video into another language with AIRecut runs an AI dubbing model, such as ElevenLabs dubbing, that translates a video's speech into another language in the same voice.