Editing with AI

LinkedIn Video: Format, Length, Captions and Prompts

Cut Express · · 5 min read

A LinkedIn video works when it makes one useful point for one kind of professional, readable with the sound off. Choose vertical for phones or horizontal for a talk or a screen, keep it as short as the point allows, put the point in the first sentence and on screen, and add captions. Platform specs change, so check LinkedIn’s current help pages for exact limits; what follows is common practice and how to get the edit done from a chat.

What makes a LinkedIn video work?

People open LinkedIn between tasks, often at a desk, often muted. That shapes the video:

  • A point in the first line. “Three things I’d change about our launch” tells a reader why to stay. A slow intro loses them (the first three seconds).
  • Captions, always. Muted viewing is common in a work setting, so the words have to be on screen.
  • One idea. A lesson, a result, a how-to or a short story. Save the second idea for the next post.
  • A real person. A face and a normal voice read as a colleague. A heavy template reads as an ad.
  • Proof. A screen, a number you can back up, a before and after, a customer’s words.
  • A calm finish. A question or one next step. The post text can carry the link.

Which format and length should you choose?

Kind of video Format that suits it Common length What to add
A tip or opinion to camera 9:16 Under a minute Captions, a title on the first frame
A product or screen walkthrough 16:9 One to two minutes Zooms on the screen, labels
A customer quote 9:16 or 16:9 Under a minute Name and role, captions
A clip from a talk or podcast 9:16 cropped on the speaker Under a minute Captions, the topic as a title
An event recap 16:9 About a minute Event name, one spoken line
A company or hiring story 16:9 One to two minutes Names, a clear close

These lengths are habits that tend to hold attention, not rules. Our post on how long a reel should be explains how to choose by goal.

How do you make a LinkedIn video in a chat?

Cut Express is a video editor you drive from a chat in Claude, ChatGPT or Claude Code. The AI looks at the footage and writes the edit; Cut Express renders the MP4. It renders 9:16 and 16:9; you post the file yourself.

  1. Upload what you have. A take to camera, a talk, a screen recording, a call with a customer.
  2. Ask for a transcript. It is made once per file, on credits, with word timings. Captions and cuts both come from it.
  3. Say who it is for and what the point is. “For heads of support: how we cut reply time.” The AI then knows which sentence to open on and what to drop.
  4. Ask for the edit. The pauses cut, the best line first, captions in a plain preset, a title on the first frame, your name and role in a lower third, the voice cleaned, and quiet music or none.
  5. Give notes. “Shorter”, “the title covers my face”, “no music”. Each note renders the next version.

A horizontal talk can be cropped on the speaker for a vertical post (horizontal to vertical). With no footage at all, a point can be made as a short piece of motion graphics: a chart, a counter, a list that builds.

What prompts can you copy?

A tip to camera:

This is a [LENGTH] take of me talking about [TOPIC] for [AUDIENCE] on
LinkedIn. Transcribe it and cut a 45-second 9:16 video: open on my clearest
sentence, cut the pauses, clean captions near the bottom, the title
"[HOOK LINE]" for the first three seconds, and a lower third
"[NAME], [ROLE]". Clean the voice. No music.

A clip from a talk:

From the talk I uploaded, find the three strongest 30 to 50 second passages
for [AUDIENCE] and show me each with its first and last sentence.
After I choose, cut that one in 9:16 cropped on me, with captions and the
topic as a title on the first frame.

A point with no footage:

Make a 20-second 16:9 motion graphics video for LinkedIn. The point:
[ONE SENTENCE]. Show it as [A BAR CHART / A LIST OF THREE / A BEFORE AND
AFTER] with these numbers: [NUMBERS]. Colors: [BRAND COLORS]. Calm,
no voice, a soft click on each item. Ask me before choosing fonts.

More in the prompt library, and the prompt guide covers the blanks.

What mistakes should you avoid?

  • No captions. The most common reason a good take gets scrolled past.
  • An intro about yourself. The lower third says who you are. Start with the point.
  • Feed tricks from other apps. Fast effects and loud tracks sit badly in a work feed.
  • Tiny screen text. Zoom the recording until a phone can read it (screen recording editing).
  • A number without a source. Colleagues and competitors are watching.
  • Captions with misheard names. Read them once before posting; product and people names are what speech recognition gets wrong.

The moves behind a good take are in talking head video editing, and client quotes are in testimonial video editing.

FAQ

Is vertical or horizontal better on LinkedIn? Both are common. Vertical fills a phone screen; horizontal suits talks and screens. Pick by what is in the picture.

Can I get a square video? Not from Cut Express. It renders 9:16 and 16:9.

Should the captions be burned into the video? Burned-in captions always show and you control how they look. Cut Express writes them into the picture, in the preset, font, color and position you ask for (caption styles).

Can Cut Express post to LinkedIn for me? No. You download the MP4 and post it.

How short is too short? If the point lands, it is long enough. A clear 20 seconds beats a padded 90.

More answers on the blog.

Make your next LinkedIn video with Cut Express

Cut Express is the video editor you drive from Claude: upload the take, say the point and the audience, get the captioned MP4. See the plans.

Edit videos by chatting with Claude.

Start your 7-day free trial

7 days free, then $19/mo. Cancel before day 7 and you pay nothing.

More in Editing with AI

7 days free, then Pro from $19/mo Start free trial

Contact support

We reply by email, usually within one business day.