AI Caption Generator for Video: Word-Timed, Styled by Chat
Cut Express works as an AI caption generator for video inside your Claude chat. Claude transcribes the speech with a timestamp for every word, then writes captions that appear as each word is spoken, in the style you ask for. Cut Express burns them into the MP4.
You don’t type a transcript or drag subtitle blocks. You describe the look, such as “bold yellow captions, two lines, centered low”, and check the render.
How does an AI caption generator work?
Auto captions come from three steps. Knowing them helps you ask for better results.
- Transcript. The speech in your clip is turned into text with timestamps. In Cut Express this runs once per file and costs 0.1 credit per minute of audio.
- Word timings. Each word gets its own start and end time. This is what lets captions change word by word instead of a full sentence sitting on screen.
- Style. The words are grouped into lines and drawn in the font and position you choose. Claude chooses the grouping from your instructions.
Claude also sees the transcript itself. That matters for editing: when Claude cuts a pause or a false start, the captions follow the cut, because both come from the same timeline.
How do I add captions to a video automatically?
- Connect Cut Express to Claude. Use the setup guide.
- Upload your video. Use the upload panel in chat, or upload from Claude Code for local files.
- Ask for captions and a style. Name a preset or describe it.
- Say how many words per line. Short-form usually reads better with 2 to 4 words at a time.
- Render, then correct. If a name is spelled wrong, tell Claude the right spelling.
Transcribe this clip and add word-by-word captions. Pop style, white with the current word in yellow, 3 words per line, lower third of the frame. Spell the brand name “Cut Express”.
Which caption styles can I choose?
Cut Express has six caption presets. You can change font, colors, size and position on top of any of them.
| Style | What it looks like | Good for |
|---|---|---|
| Pop | Words pop in as spoken | Energetic reels and Shorts |
| Karaoke | The line stays, the spoken word changes color | Following along, lyrics, tutorials |
| Boxed | Words on a solid box | Busy or bright backgrounds |
| Bold | Heavy type, high contrast | Hooks and punchy talking heads |
| Clean | Simple, readable text | YouTube videos and explainers |
| Minimal | Small, light captions | Calm or cinematic edits |
For a deeper look at timing, read word-by-word captions. For vertical video, see captions for Reels. Karaoke has its own guide: karaoke captions.
Can I caption an AI voice-over?
Yes, and you skip the transcript. When Claude generates a voice-over from your script, the voice comes back with word timings. Claude uses them for the captions directly. See AI voiceover for video.
What does a real caption request look like?
Say you have a 60-second talking-head clip for YouTube. A first pass could go like this:
Make this 16:9. Cut the long pauses, then add clean captions at the bottom, max 6 words per line. Put the title “3 mistakes with LUTs” as a text box in the first 3 seconds.
After the render you might notice the captions cover your hands in one shot. You’d follow up:
In the scene at the 0:40 mark, move the captions to the top.
Claude changes that one part of the timeline and renders the next version. Earlier versions stay available.
What should I still check?
Auto captions are only as good as the transcript. Check these by eye:
- Names and brand terms. Tell Claude the spelling up front.
- Overlaps. Captions over a face or a product label may need a new position.
- Line length. If lines feel crowded, ask for fewer words per line.
When is another caption tool a better fit?
Cut Express renders captions into the video itself. That suits short-form video, where the captions are part of the look. It’s less suited to a few cases:
- Viewers who want to turn captions off. A platform’s own closed-caption upload lets them toggle captions. Burned-in text can’t be hidden.
- Several languages on one video. Separate subtitle tracks per language are easier to manage in a platform’s caption settings or a dedicated subtitle tool.
- Long transcripts you need to proofread line by line. A subtitle editor with a text view is quicker for heavy correction work.
If you need any of those, use a subtitle tool for the track and Cut Express for the edit.
FAQ
How much do AI captions cost in Cut Express?
The transcript costs 0.1 credit per minute of audio, once per file. Captions from an AI voice-over need no transcript. Pro includes 200 credits a month.
Are the captions burned in or a separate file?
They are rendered into the MP4.
Can I add titles and labels as well as captions?
Yes. Free text boxes handle titles, labels and calls to action, placed wherever you ask.
Does it work for vertical and horizontal video?
Yes. Cut Express renders 9:16 and 16:9, and Claude can make one from the other on the same timeline.
Caption your next video
Upload a clip and describe the caption look you want.
Free for 14 days, then Pro is $19/month. Cancel before the trial ends and nothing is charged.