How to Write a Short-Form Video Script People Actually Finish

Most founders film short-form video the same way: prop the phone against a laptop, hit record, and start talking. Thirty seconds in, they've circled the same point twice, said "basically" four times, and still haven't told anyone what the video is about. Take six is better than take one, but it's still a conversation looking for a structure.

The fix isn't better lighting or a nicer mic. It's writing the thing down first.

A short-form script isn't a teleprompter monologue, and it isn't a screenplay. For a 30-second video it's about four lines on a page. Writing those four lines takes five minutes and saves you twenty — because you stop re-recording to find the point you should have decided before you pressed record.

A script is four lines, not four paragraphs

Here's the shape that works for almost any talking-head, demo, or tip video:

1. Hook — the promise. Why keep watching?
2. Context — the stakes. Why does this matter to you specifically?
3. Payoff — the one idea, delivered concretely.
4. Close — one action, or one clean ending.

That's it. Four lines. If you can't fit your idea into them, you don't have one video — you have two or three, which is good news for your calendar.

The common failure is spending the whole runtime on step 2. Founders love context. Viewers don't — they want the promise kept quickly.

Line 1: The hook is a promise, not a greeting

Cut "Hey guys, welcome back." Cut the logo sting. Cut the slow zoom while you smile. A short video has roughly a second and a half before someone decides, and a greeting spends it on nothing.

A hook works when it does one of three things:

• States a specific outcome: "This is how we plan a month of posts in one sitting."
• Names a problem out loud: "Your posts aren't flopping because of the algorithm. They're flopping because of the first line."
• Breaks an expectation: "We stopped posting daily. Reach went up."

Write the hook last if it helps — it's usually the sharpest sentence hiding somewhere in your payoff. Pull it to the front.

One rule that saves most videos: the hook should be understandable with the sound off. Say it out loud and put it on screen as text.

Line 2: Earn the next five seconds

This is one sentence, not a backstory. It tells the viewer who this is for and why it's worth their time.

"If you're a founder doing your own social on a Sunday night, this is the part that was eating my whole evening."

That's the entire job. You're narrowing the audience on purpose — the person who recognizes themselves stays, and everyone else was never going to convert anyway.

Line 3: The payoff — one idea, fully delivered

The mistake here is width. A 30-second video with three tips delivers none of them; a 30-second video with one tip, shown properly, is useful.

Make it concrete. "Batch your content" is a slogan. "Open your calendar, block 90 minutes on Monday, and write every caption for the week before you open a single design tool" is a thing someone can do today.

If you're demoing something, show the actual screen, not a description of the screen. If you're making a claim, give the specific example behind it. Specificity is what makes a short video feel like it came from someone who has actually done the work.

Line 4: The close — one action

Pick exactly one: follow, comment a word, read the full version, try the thing. Two calls to action is the same as none.

And "comment X and I'll send it" only works if you can actually handle the replies. If a video does well and forty people comment, those forty comments are now a job — one that lives in your inbox, not your calendar. Make sure the ask matches the capacity you have to answer it.

Write it for the ear, then cut 30%

Read the script out loud. Anything you stumble over is written, not spoken — rewrite it the way you'd say it to a friend. Contractions are good. Sentences under twelve words are good. Subordinate clauses are where founders go to die on camera.

Then cut roughly a third. Almost every first draft has a warm-up sentence at the top, a clarifying aside in the middle, and a trailing "so yeah" at the end. None of them survive contact with a scroll.

The column most people skip

Write your script in two columns: what you say on the left, what's on screen on the right.

The right-hand column is where short-form video is actually won. It holds the on-screen text for the hook, the cut points, the B-roll or screen recording, the product shot, the caption styling. Deciding it before you film means you record the right footage once instead of improvising coverage later.

It also forces you to notice when the visual is just you talking for 30 straight seconds — which is fine occasionally, and deadly as a habit.

Where AI genuinely helps (and where it doesn't)

AI is good at the mechanical parts of scripting: turning a rough idea into five hook variations so you can pick the sharpest, tightening a rambling draft to a target runtime, rewriting the same script for a 15-second cut and a 45-second cut, and producing the caption and on-screen text that go with it. That's real time saved, and it's the kind of work where generating six options beats agonizing over one.

What AI can't do is supply the specific. It doesn't know the number you saw last quarter, the thing a customer said on a call, or the mistake you made in year one — and those are precisely the details that make a video worth watching. Use the model for structure and variation; bring the substance yourself.

In Trendly this sits in one place rather than four tabs: generate the script, the caption, and the visual direction from a single prompt, have a teammate review and approve before anything goes out, then schedule it to every platform from the same calendar. The point isn't that AI writes the video — it's that the gap between "idea" and "scheduled" stops being an evening's work.

The part no script can fix

A script fixes rambling. It doesn't fix having nothing to say.

If you sit down to write four lines and can't name the one idea, that's useful information: the video isn't ready, and no amount of filming will make it ready. Go back to your idea bank, pick something you have an actual opinion or example for, and script that instead.

The founders who are consistent on video aren't more comfortable on camera. They've just stopped treating every post as a performance and started treating it as four lines they wrote before lunch.