AI Voiceover for Video: Script to Voice to Captions
To add an AI voiceover for video with Cut Express, you paste a script into Claude and ask for a voice-over. Claude generates a natural AI voice, lines your footage up under it, times captions to each spoken word and lowers the music while the voice plays.
It works for 16:9 YouTube videos and product demos as well as vertical reels. If you only make Shorts, our guide to AI voice-over for Shorts covers that case. This post is about everything else.
What does an AI voiceover for video actually involve?
A voice-over is more than an audio file. For it to work in a finished video, four things have to line up:
- The script becomes speech with a natural voice.
- The footage has to match what the voice is saying at each moment.
- The captions have to appear as each word is spoken.
- The music has to drop when the voice talks and come back in the gaps.
In a timeline editor you do steps 2 to 4 by hand. With Cut Express, Claude does them in one pass. The voice comes back with word timings, so Claude knows the exact second each word starts. It uses those timings to place cuts and captions. Music ducks under the voice by itself.
How do I add an AI voiceover to a YouTube video?
Here is the full flow for a 16:9 video:
- Connect Cut Express to Claude. Follow the setup guide. It works in Claude.ai and Claude Desktop, and in Claude Code for local files.
- Upload your footage. Drop your screen recordings and b-roll into the upload panel. See how to upload footage.
- Paste your script. Plain text is fine. Mark any lines that must sit over a specific clip.
- Ask for the edit. Tell Claude the format, the tone of the voice and which clips go where.
- Watch the render and give notes. Claude edits the same timeline and renders the next version.
A first request can be this short:
Here’s my script for a 3-minute YouTube explainer. Make a calm, clear voice-over from it, 16:9. Use the screen recordings for the steps and the b-roll for the intro. Captions at the bottom, clean style. Add the music track under it.
Claude watches each clip through contact sheets (frames labelled with source time). It matches each sentence to footage that shows it. For example, the line “open the settings page” lands on the clip where the settings page opens.
Can I use AI voiceover for a product demo?
Yes. Product demos are a good fit because the voice has to describe what the screen shows, step by step. Claude can see the screen recording, so it can trim dead time between clicks and keep each step on screen while the voice explains it.
You can also add what demos usually need:
- Arrows and labels pointing at a button.
- Punch-in zooms on the part of the screen that matters.
- Lower thirds with the feature name.
- Picture-in-picture if you recorded your face too.
A demo prompt might look like this:
Script below. Make a voice-over for a 90-second product demo, 16:9. Zoom in when the voice mentions the export button and add an arrow pointing at it. Word-by-word captions, boxed style.
How do captions follow the AI voice?
The voice comes back with a timestamp for every word. Claude passes those timings straight to the captions, so each word shows up as it is spoken. You don’t transcribe anything, and nothing drifts out of sync.
You choose the look: pop, karaoke, boxed, bold, clean or minimal, plus font, color, size and position. For more on styles, read word-by-word captions.
Does the music get quieter under the voice?
Yes. Music ducks under the voice by itself and comes back up in the pauses. You can still ask for changes, such as “music a bit louder in the intro” or “fade the music out over the last 5 seconds”. Our post on ducking music under a voice explains the levels.
AI voiceover vs recording your own voice
| AI voice-over in Cut Express | Recording your own voice | |
|---|---|---|
| Setup | A script and a chat | Mic, quiet room, retakes |
| Changing a line | Edit the script, Claude re-renders | Record the line again and re-sync |
| Captions | Timed from the voice’s word timings | Need a transcript first |
| Best for | Explainers, demos, faceless channels | Personal channels where your voice is the brand |
If your audience knows your voice, keep recording it. Cut Express can transcribe your own narration and caption it the same way. If you want to move faster or run a faceless channel, an AI voice removes the recording step. See editing for faceless YouTube channels.
How much does an AI voiceover cost?
Voice-over uses credits. One credit covers 1,000 characters of script, roughly one minute of speech, or 2 credits with the Advanced voice. Pro includes 200 credits a month and Ultra includes 500. Unused credits roll over for 3 months. Render minutes are counted separately, by the length of the video you render.
FAQ
Can I change one line without redoing the whole voice-over?
Yes. Edit the line in the script and ask Claude to update it. Every version is kept, so you can go back.
Does it work for long videos?
Pro renders videos up to 10 minutes and Ultra up to 30 minutes.
Can I add sound effects as well as a voice?
Yes. Describe the sound or use the built-in free sound effects. See sound effects for YouTube videos.
What resolution do I get?
1080p on Pro, and 2K (2560×1440 for 16:9) on Ultra. The file is an MP4 (H.264).
Try it on your next video
Write the script, drop in your clips and let Claude build the edit around the voice.
Free for 14 days, then Pro is $19/month. Cancel before the trial ends and nothing is charged.