Add Captions and Subtitles

Auto-generate captions from any video's speech for 2 credits, then edit the words, pick a style preset, and change colors.

Written By Andrei Terteci

Last updated 7 days ago

Most people watch social video with the sound off. Captions keep them watching, make your video accessible, and give word-by-word text something to animate. Krumzi generates captions from your footage automatically and lets you edit every word afterwards.

Generating captions

  1. Add a video element that contains speech (uploaded footage or a clip with a voiceover).

  2. Open the Captions panel from the side navigation.

  3. Under Select Video, pick the clip you want to transcribe. Clips that already have captions are marked.

  4. Click Generate Captions.

Krumzi transcribes the audio and creates a caption element timed to the speech. Transcription costs 2 credits per file, charged only once a transcript comes back. If it fails, you're not charged.

The transcriber handles ten languages: English, German, French, Spanish, Portuguese, Turkish, Hindi, Mandarin, Japanese, and Korean. Files up to about 25 MB per transcription work best. If your clip is bigger, trim it or compress it first.

Editing the words

Select the caption element (on the canvas or the timeline). The panel opens on the Edit Captions tab, where you can fix any word the transcriber got wrong. Timing is kept, so you only touch the text.

Styling

Switch to the Style & Presets tab:

  • Caption Presets. Default, Neon Glow, and Dynamic Highlight (the word-by-word highlight style you see across TikTok and Reels).

  • Colors. Set the text and highlight colors.

  • Uppercase. Toggle all-caps.

Beyond that, a caption is a text element. Font, size, weight, and position all work the usual way. Drag it wherever it reads best. Bottom-center is the classic spot; higher up is safer when a platform's UI covers the bottom of the frame.

Timing

Captions are timed to the speech in the source clip. If you move or trim the video element, keep the caption's bar aligned with it. Generating captions again for the same clip replaces the previous ones.

Captions vs. titles

Captions transcribe what's said and change throughout the video. Titles are design elements: bigger, stylised, on screen longer. Use both. Big title text for the key message, captions for the narration.

Ask the AI

You can also ask the AI chat in the editor to add captions to a clip and pick a style. Same 2-credit transcription cost, same editable result.

Tips

  • Speak clearly and keep music low in the source clip. Cleaner audio means fewer corrections.

  • Read through the transcript once. Names and product terms are the usual misses.

  • Choose high-contrast colors. White text with a bright highlight color reads on almost any footage.

  • Check the preview on a phone-sized canvas. Captions that look fine on a monitor are often too small on mobile.