How to Write a Video Script (With Free Template)

Most video scripts don't fail because the writing is bad. They fail because the first three seconds ask the viewer to be patient, and nobody is. This guide covers the structure that holds attention, how to write words that sound right when spoken aloud, a template you can copy, and what to do with the script once it's written.

Decide one thing before you write a word

Before drafting, finish this sentence: "After watching, one specific person will be able to ___." Not "will know about" — will be able to. A script with a verb at the end of that sentence has a natural shape. A script without one wanders, and wandering is what viewers leave.

This is also the fastest way to cut. Every paragraph you write either moves the viewer toward that verb or it doesn't. The ones that don't come out, no matter how much you like them.

The structure: open, promise, body, payoff

Nearly every video that holds attention uses the same four beats:

1. Open (0:00–0:03). A statement that creates a gap the viewer wants closed. Not a greeting, not your name, not "in today's video."
2. Promise (0:03–0:15). Tell them exactly what they'll walk away with, and roughly how long it takes. This is the contract. Keep it.
3. Body. Three to five beats, each one idea. Order them so each beat depends on the one before it — that dependency is what makes people keep watching.
4. Payoff. Close the gap you opened. Then one call to action, singular. Two calls to action is the same as none.

The first 3 seconds (this is most of the job)

Retention graphs almost always show the same cliff in the first few seconds. Whatever you put there does more work than the rest of the script combined. Openers that reliably survive that cliff:

• The contradiction. "Everyone says to X. On a small channel, X is why you stall."
• The specific number. "I spent 40 hours editing last month. 31 of them were on one thing."
• The stakes. "If you get this wrong, the rest doesn't matter."
• The direct question they're already asking. "Why does your footage look flat?"

Openers that don't work: greetings, self-introductions, "in this video I'm going to," and any sentence whose first four words could belong to any video on any topic.

Write for the ear, not the page

Written prose and spoken prose have different rules. Sentences that read fine on screen turn into breath problems when narrated. Four habits fix most of it:

• One idea per sentence. If there's a comma doing structural work, it's probably two sentences.
• Use contractions. "You'll" and "it's" are how people talk. "You will" reads as a manual.
• Put the subject early. Front-loaded clauses make the listener hold state until the verb arrives.
• Say numbers the way you'd say them. Write "about a third" if that's what you'd say out loud.

The test is not re-reading. It's reading aloud, at pace, with a timer. Anywhere you stumble or run out of breath, cut the sentence in half. This one pass improves a script more than any amount of silent editing.

How long should it be?

Count words, not pages. Conversational narration lands around 140–160 words per minute, so:

• 60-second short → ~150 words
• 5-minute explainer → ~750 words
• 10-minute deep dive → ~1,500 words

Then ignore the estimate and time your read-aloud. Pauses, emphasis and breathing all cost seconds that word counts don't capture.

A template you can copy

TITLE:
ONE-SENTENCE GOAL: After watching, [who] will be able to [verb].

[0:00] OPEN — the gap
  One sentence. No greeting. No name.

[0:03] PROMISE — the contract
  What they get + roughly how long.

[0:15] BEAT 1 — [one idea]
  Claim. Evidence. Consequence.

[BEAT 2] — [one idea, depends on beat 1]
  Claim. Evidence. Consequence.

[BEAT 3] — [one idea, depends on beat 2]
  Claim. Evidence. Consequence.

[PAYOFF] — close the gap you opened
  Restate the goal, now answered.

[CTA] — exactly one action.

Keep the bracketed timestamps while drafting. They're a budget: if beat 2 has crept to 3 minutes, you'll see it before you record instead of after.

The second pass: what's on screen

Writing and visualizing at the same time slows both down. Draft the words first. Then make one pass down the script and mark, line by line, what should be visible: the number you just quoted, the term you just defined, the three steps you just listed.

This pass is where most of the production time actually goes. Searching a stock library for a clip that vaguely matches "compound interest grows on itself" can take longer than writing the whole script did — and the clip still won't be about your point.

There's a shortcut worth knowing: once your narration is recorded and you have subtitles, the subtitle file already contains both what you said and exactly when. That's the full cut list. Ahacut reads it and generates a full-length motion-graphics track — animated text, numbers, charts — timed to every line, which you then drop onto your timeline in CapCut, Premiere or Final Cut. Nothing to switch, nothing to nudge into place.

Five mistakes that cost you viewers

1. Warming up on camera. Whatever you say before the first real sentence should be cut. Start the script at the point where the value starts.
2. Burying the promise. If viewers can't tell by 0:15 what they're getting, they leave — even if the payoff is excellent.
3. Two ideas in one beat. The second one gets lost. Split it or drop it.
4. Writing to impress. Long words and long sentences read as effort, not authority.
5. Ending with three CTAs. Like, subscribe, comment, join, download — pick one. The others cancel it out.

Frequently asked questions

Related: How to add b-roll to videos automatically · What is b-roll? A-roll vs b-roll · How to write a transcript of a video