How to Write a Voiceover Script (Format and Example)

Published November 8, 2026 · Updated November 8, 2026 · 6 min read · by the Sakura5 team

To write a voiceover script, write for the ear instead of the eye: short sentences, one idea per line, words matched to what's on screen at that moment, and about 150 words per minute of finished audio. Then read it aloud, mark where you breathe, and cut anything that trips your tongue before you record.

A voiceover is different from a talking-head script because nobody sees your face. The audience can't read your expression to catch your meaning, so the words and your delivery carry everything. It also has to fit the picture — a product demo, a slideshow, an explainer animation, a travel montage — which means timing matters more than it does when you're on camera.

Below: the format that works for voiceover, a step-by-step method, a worked example for a 45-second app demo, and how to record clean audio on an iPhone while reading the script.

What Makes a Voiceover Script Different

Three constraints shape every voiceover:

• The picture sets the clock. If the screen shows the settings menu for six seconds, you have six seconds to talk about the settings menu. Write to the visuals, not the other way around.
• Listeners can't re-read. A reader can glance back at a long sentence. A listener hears it once. Anything with nested clauses, acronyms or a list of more than three items gets lost.
• Breath is part of the script. A sentence you can't say in one breath will sound strained. Line breaks are where you breathe.

That's why voiceover scripts look short and choppy on the page. That's correct. Smooth prose on paper usually sounds stiff out loud.

The Voiceover Script Format

Professional voiceover scripts often use a two-column layout: visuals on the left, narration on the right, lined up by scene. You don't need special software for this. A simple version works in any notes app:

• Scene label — what's on screen, written for you, not to be read aloud.
• Narration — the exact words, one sentence per line.
• Timing — how many seconds the scene lasts.

When it's time to record, strip the scene labels out and keep only the narration. Anything left on your reading screen is something you might read aloud by accident, so the copy you record from should contain nothing but spoken words.

How to Write a Voiceover Script, Step by Step

1. Lay out the visuals first

List the scenes in order with their length in seconds. If the video is already edited, scrub through it and write down the timestamps. If it isn't, plan the visuals before the words.

2. Budget words per scene

At about 150 words per minute, you get roughly two and a half words per second. A five-second scene holds about twelve words. A three-second transition holds a short phrase or nothing at all. Silence over a strong visual is allowed.

3. Write one sentence per idea

Say what the viewer is looking at, then why it matters. "This is the dashboard. Everything you scheduled today is here." Two sentences, two breaths, both easy to hear.

4. Use spoken words, not written ones

Contractions, plain verbs, numbers written the way you'd say them. "Twenty-five percent" is easier to read correctly than "25%." Spell out anything you might hesitate on, and write tricky names phonetically if you need to.

5. Mark pauses

Put a line break where you want a beat. An ellipsis or a blank line works too. These marks are what make the read sound natural rather than rushed.

6. Read it against the picture

Play the video and read along, out loud. Wherever you run long, cut words; wherever there's dead air you don't want, add a short line. Two passes usually get it tight.

A Worked Example: A 45-Second App Demo

Here's a voiceover for a short demo of a habit-tracking app, laid out by scene:

1. Scene 1, 0–5s, app icon on home screen: "Most habit trackers want an hour of setup. This one takes a minute."
2. Scene 2, 5–15s, adding a habit: "Type the habit. Pick the days. That's it — it's on your list for tomorrow morning."
3. Scene 3, 15–28s, checking off habits: "Each morning, you tap what you did. No forms, no notes unless you want them. The ring fills as the day goes."
4. Scene 4, 28–40s, weekly view: "At the end of the week, you see the pattern. Which days you showed up. Which habit keeps slipping."
5. Scene 5, 40–45s, logo: "Start with one habit. Add the next one when it sticks."

That's about 90 words for 45 seconds — a little under the 150-per-minute pace, which is right for a demo. The viewer needs time to look at the screen, and a voiceover that talks over every frame feels like a sales pitch.

Recording the Voiceover on an iPhone

You don't need a booth. You need a quiet room with soft surfaces, the phone about a hand's width or two from your mouth and slightly off to the side, and a way to read the script without rustling paper.

That last part is where a teleprompter app helps. In Cueflow you can record audio only: the script scrolls on screen at the speed you set, you read it, and you get a clean audio file to drop into your editor. Adjust the text size and scroll speed until the words arrive exactly when you need them, and the reading-time estimate tells you up front whether the script fits the 45 seconds you have.

If the voiceover is narrating something on your phone — an app walkthrough, a tutorial — Cueflow can also record your screen, so you can capture the demo and your narration in the same session. The prompter overlay never appears in the recording.

A few recording habits that make editing easier:

• Record each scene's lines as a separate take, with a second of silence before and after.
• If you flub a line, pause, then repeat the whole sentence. Cutting at a sentence boundary is invisible; cutting mid-sentence isn't.
• Smile on upbeat lines. It's audible.

Using AI for the First Draft

If the blank page is the slow part, an AI script writer can produce the first version. In Cueflow you describe the video, pick a tone and an exact length, and get a draft written as clean spoken words, with no markdown or stage directions to read by mistake. Then do the steps above: lay it against your scenes, cut to your word budget, and use the rewrite options to make a line shorter or fix the grammar. For very short pieces, the same approach is covered in our guide to writing a 60-second video script.

FAQ

How many words is a one-minute voiceover?

Around 150 words is a common working figure for a one-minute voiceover at a natural pace. Explainers and demos often run slower, around 120–140, so viewers can follow the screen. Time yourself reading a paragraph aloud and use your own pace as the budget.

How do you format a voiceover script?

The standard format pairs each scene's visuals with the narration for that scene, plus its length in seconds. Write one sentence per line, use line breaks for pauses, and strip the scene notes out of the copy you actually read from so you never say them aloud.

Can I record a professional voiceover on my iPhone?

Yes, for most online video. Record in a quiet room with soft furnishings, keep the phone close to your mouth but slightly off-axis to avoid pops, and read from a scrolling script rather than paper. Record each scene as a separate take to make editing easy.

Should a voiceover script be read word for word?

Mostly, yes. Voiceovers have to fit the picture, so improvising tends to run long. Read the script as written, but if a line feels unnatural when you say it, change the wording in the script before recording rather than paraphrasing on the fly.

← All articles