How to write a video script (format, example and template)
Most advice on how to write a video script stops at “know your audience” and “have a strong hook.” True, and useless when you’re staring at a blank page with a shoot on Thursday.
I’ve edited over a thousand videos. The scripts that cut together easily all had the same shape, and the ones that didn’t were missing the same things. This is the process I’d hand you if you were sending me footage: a format, six steps, and a finished example you can copy.
What a video script looks like
A video script is what the viewer hears and sees, in order, with how long each part lasts. That’s it. If a document doesn’t tell you all three, it’s notes, not a script.
The format that does this best is the two-column A/V script, split into scenes. Each row is one scene. The words go on one side, the picture on the other, and the length sits at the end so you can add it up.
| Scene | What they hear | What they see | Length |
|---|---|---|---|
| 1. Hook | “This mistake cost me a full day of shooting.” | Close-up, me holding an empty SD card | 0:05 |
| 2. Stakes | “If you shoot on more than one card, it can happen to you too.” | Desk with three cards and a laptop | 0:06 |
| 3. Fix | “Here’s the two-minute habit that stops it.” | Screen recording of the folder setup | 0:08 |
You can write this in a doc, a spreadsheet or a dedicated tool. What matters is that every line of narration has a picture next to it and every scene has a length.
How to write a video script in 6 steps
1. Write the promise in one sentence
Before any outline, write what the viewer will have or know by the end. “By the end, you’ll have a file naming system that sorts itself.” If you can’t write that sentence, you don’t have a video yet. Every scene you write later gets checked against it.
2. Outline the scenes before writing lines
Don’t start with the first sentence. Start with the scene list:
- Hook: the first few seconds, the reason to keep watching.
- Stakes: why this matters to the viewer, or what goes wrong without it.
- Body: 3 to 5 scenes, each making one point.
- Payoff: the result, shown, not just described.
- CTA: one ask, not three.
Give each scene a working title and a rough length. Now you can see the whole video before you’ve written a word of it, which is where most structural problems get caught cheaply. I go deeper on this in scene-based scriptwriting.
3. Write each scene for the ear
A script is read aloud, not read. Write every scene, then say it out loud. If you stumble, the viewer will too.
- Keep sentences short. One idea each.
- Write numbers the way you’ll say them: “about two thousand” rather than “~2,000”.
- Cut words you’d never say in conversation, like “furthermore” or “utilize”.
4. Fill in what the viewer sees, line by line
Now go back to the right-hand column and fill it for every line. A shot, a screen recording, a graphic, a piece of B-roll, your face.
If you can’t think of a visual for a line, the line is usually too abstract. “Organization is important” has no picture. “Three cards, one of them unlabeled” does. Rewrite the line until it has something to show. A storyboard template helps here if you think in frames.
5. Check the length
A common working average for narration is about 150 words per minute, with most people landing somewhere between 130 and 170. Use it to check each scene against its target, not just the total.
| Video length | Words at ~150 wpm | Range (130–170 wpm) |
|---|---|---|
| 1 minute | 150 | 130–170 |
| 5 minutes | 750 | 650–850 |
| 8 minutes | 1,200 | 1,040–1,360 |
| 10 minutes | 1,500 | 1,300–1,700 |
| 15 minutes | 2,250 | 1,950–2,550 |
Paste your scene into the script length calculator for a quick read, and see how many words a 10-minute YouTube video needs for what pushes those numbers up or down.
6. Mark what can be cut
Every script runs long once it’s on a timeline. Before you shoot, mark the lines and scenes that could go without breaking the promise. In the edit, that’s the first place I look. It saves an argument with yourself at 1 a.m. and keeps the video from losing a scene you actually needed.
A short example
Here’s a complete 60-second script in the same format. Topic: how to name your video files so you never lose a take.
| Scene | What they hear | What they see | Length |
|---|---|---|---|
| 1. Hook | “You shot the perfect take. Now it’s called C0047, and so are three others.” | Finder window full of identical camera file names | 0:06 |
| 2. Stakes | “I’ve watched editors lose good takes this way. Not deleted. Just buried in a folder of names nobody can read.” | Slow scroll through a messy footage folder | 0:08 |
| 3. The format | “Here’s the format I use. Date first, year then month then day. Then the project, the scene number and the take. The files sort themselves.” | Text builds on screen: 2026-10-13_Kitchen_S03_T02 | 0:14 |
| 4. When to rename | “Rename on the day you shoot, before the card goes back in the camera. Tomorrow you won’t remember which C0047 was the good one.” | Hands swapping an SD card, laptop renaming files | 0:12 |
| 5. Mark the keeper | “Then add one word to the best take. Keep, or star. Pick one and never change it, so a search finds every keeper you’ve ever shot.” | Search bar typing “keep”, results filter down | 0:12 |
| 6. CTA | “Do it on your next shoot. If it saves you one search, follow for the next editing habit.” | Back to camera, end card | 0:08 |
Six scenes, one promise, a picture for every line, and the lengths add up to a minute. Notice that scene 3 is the longest: it carries the actual information, so it gets the time.
YouTube scripts vs other video scripts
The format above works for any video, but YouTube changes a few things.
The hook has to land in the first seconds. A brand video plays to a captive audience. A YouTube viewer can leave with one swipe. Put the promise, or the problem, in the first scene.
Plan chapters in the script. If your scenes have clear titles, they double as chapter timestamps in the description. Viewers use them to jump around, and you’ll know where they go before you edit.
Write retention beats. Every 30 to 60 seconds, give the viewer a reason to stay: a reveal, a change of visual, a “here’s where it gets interesting.” In a scene-based script, these usually sit at the start of each body scene.
Shorts are their own thing. A Short is vertical and usually only a few scenes long. The structure compresses to hook, one point and payoff, and the visual column matters even more because most of it is framed tight.
If you want this ready to fill in, grab the free YouTube script template.
How I do this in Syntale
I built Syntale around this format because I kept rebuilding it in docs. A script is split into scenes, and each scene has its own duration and a B-roll and visual plan sitting next to the words. Hover a scene marker and you see its length, like “0:06”, so step 5 happens while you write. You can keep a few versions of a tricky paragraph and switch between them, share the project for comments and approval, and export a PDF your editor can cut from. Start for free or use the in-app template.
FAQ
What does a video script look like?
Usually a two-column table: what the viewer hears on one side, what they see on the other, split into scenes with a length for each. Narration-only scripts exist, but they leave the picture to guesswork.
How long should a video script be?
At about 150 words per minute, a 5-minute video is roughly 750 words and a 10-minute video roughly 1,500. Your own pace may sit anywhere from 130 to 170, so time yourself and adjust.
Can ChatGPT write a YouTube script?
It can draft one, suggest structures and give you alternative lines when you’re stuck. It doesn’t know what you’ll actually show on screen or how you talk, so use it for options and write the final version yourself.
What’s the best way to write a YouTube script: Docs, Notion or a dedicated tool?
Google Docs is fine for short, narration-led scripts, and Notion works if you already plan everything there. Both make you build the scene structure and timing by hand, which is the gap a scene-based tool like Syntale fills. I compare them in detail in Google Docs vs Notion vs Syntale.