How to Add a Voiceover and Word-by-Word Subtitles Without Recording Your Voice

Plenty of people who would make videos do not want to hear their own voice on them, or do not have a quiet room, a microphone, or the patience for twenty takes. A natural AI voice solves the recording. What makes the result feel professional is everything around it: a script written to be spoken, subtitles that keep pace with the voice, and sound under it.

Here is how the agent handles each part, and what you can control.

1. A script written for the ear

Writing for reading and writing for listening are different. Spoken scripts use shorter sentences, one idea at a time, and words that are easy to say aloud. When you ask for a narrated video, the agent writes the narration in that style and splits it into parts, one per scene.

Then it shows you the whole narration and waits. You can change any line, cut one, or write your own. Nothing is recorded until you approve it, so editing the script is free and fast.

2. Choosing a voice

There are dozens of natural voices across 9 languages: warm, authoritative, friendly, calm; male and female. Rather than making you scroll a list, the agent suggests a few that suit your topic and language, each with a short description, and recommends one. You can also name the voice you want, or ask for a different one at any time.

Each part of the narration is recorded separately. That has a practical benefit: when you change one line later, only that line is recorded again, and the rest of the video stays exactly as it was.

3. Subtitles that follow the voice

Most viewers on social media watch with the sound off, so subtitles are not optional. Typing them out and nudging each one into place is the slowest part of editing.

Here there is nothing to type. Subtitles are generated from the narration itself and timed to the voice, so each line appears as it is spoken. The on-screen titles in each scene are timed to the narration too: a number appears on the word that says it.

slammed into what is now Mexico.

11.20 / 66.4s

Subtitles Ask about this frame Frame Export MP4

The subtitles switch in the preview bar turns them on or off, and the exported MP4 follows it. Turn them off if the platform adds its own captions, or if you plan to add styled captions elsewhere. The agent keeps the subtitle area at the bottom of the frame clear of other text, and the preview checks every frame to make sure nothing sits on top of them.

4. Music and sound effects

Silence under a voice sounds unfinished. The agent can add a quiet background bed (ambient, lo-fi or electronic pulse) mixed below the narration so it never competes with it, and short sound effects that land on moments in the video: a whoosh as a title enters, an impact as a picture lands.

Prefer no music? Say so. Want it louder, quieter, or a different mood? Ask.

5. The same video in another language

Your script and voice can be in English, Simplified or Traditional Chinese, Japanese, Korean, French, German, Spanish or Italian. To reach another audience with a video you have already made, ask for a version in that language: the agent translates the narration and the on-screen text, shows you the translated script to approve, and records it with a voice in that language. The scenes stay the same.

When to skip the voice entirely

Not every video needs narration. Short, punchy social videos often work better as text and music only: a list, a quote, a set of numbers. On the first question form, choose “No voiceover” and the agent designs the video to be read rather than heard, with the text on screen long enough to read.

A quick checklist before you export

• Listen once with your eyes closed. Does the narration make sense without the pictures?
• Watch once on mute. Can you follow it from the subtitles and titles alone?
• Check names and numbers in the subtitles. They are what viewers screenshot.
• Decide on subtitles on or off for the platform, then export.