How to turn long videos into shorts without watching them back

If you already publish long videos, you are sitting on the cheapest short-form content you will ever make. It is already recorded, already yours, and already proven to hold attention for the people who watched it.

The problem is finding the good parts. An hour of footage might contain six clip-worthy moments, and locating them means watching the whole thing again with a notepad.

Why the transcript is the shortcut

You do not need to watch a video to know where the good moments are. You need to read it. A timestamped transcript makes the whole episode searchable, and a strong moment reads as strong on the page: a claim, a reversal, a number, a story that resolves.

This is also why cost follows the source video rather than the number of clips. The transcript is the expensive part, and it is the same transcript whether you cut one clip from it or ten.

Practical consequence: cutting ten clips from one long video costs roughly the same as cutting one. If you are clipping at all, clip properly.

What makes a moment clip-worthy

A clip has to work for someone who has no idea what came before it. That rules out most of what feels good in context.

• It stands alone. No "as I was saying" and no unresolved pronouns.
• It starts at the claim, not at the wind-up to the claim.
• It ends at the resolution, not four seconds after it.
• It has a reason to exist: something surprising, useful, or contested.

The most common failure is starting too early. The speaker takes eight seconds to arrive at the point, and those eight seconds are exactly where a scrolling viewer leaves.

Vertical framing is not just cropping

A 16:9 frame cropped to 9:16 loses two thirds of the width. If the speaker is off-centre, which they usually are, a naive crop puts them half out of frame.

Framing has to follow whoever is talking. On a two-person interview that means the crop moves between them as the conversation does. Get this wrong and the clip is unwatchable no matter how good the moment was.

Captions are not optional

A large share of short-form viewing happens with sound off. A clip without captions is a clip that most of the feed will scroll past without ever knowing what it was about.

Word-level timing matters more than styling here. Captions that lag the audio by even a moment feel broken in a way viewers notice without being able to say why.

What still needs you

Automation is good at finding candidates and bad at knowing what you would be embarrassed to post. Two things are worth checking before anything goes out:

• Context that reads badly alone. A joke that lands in a forty minute conversation can look very different as a standalone clip.
• Rights. Clipping a channel you do not own is a licensing question, not a technical one.

A workflow that holds up

• Track the source channel so new uploads are picked up without you.
• Let the transcript pick candidate moments and cut them vertically with captions.
• Skim the candidates, not the source. Reading six clip titles takes a minute.
• Publish on a spacing schedule rather than all at once, so one video does not flood a feed.

The whole point is that the expensive step, deciding what is worth clipping, stops requiring your attention. Reviewing the output still should.

Common questions