How to Turn a Long Video Into Shorts, Step by Step

A podcast episode, a webinar, a lesson or a stream usually holds a handful of moments that work on their own as vertical shorts. Getting them out is the same job by hand or with a tool.

The short version, in seven steps:

1. Find the moments. Skim the transcript, not the video, and mark passages that make one complete point without needing what came before.
2. Cut each clip on the thought. Start on the first line that stands alone and end on the payoff. Cut throat-clearing at the start.
3. Reframe to 9:16. Crop the landscape picture to vertical (1080×1920) and keep whoever is speaking in frame.
4. Add captions. Burn in word-timed captions and fix the names and jargon that transcription gets wrong.
5. Make the first seconds count. Open on the strongest line, or add a short on-screen hook.
6. Export one vertical file. An MP4 at 1080×1920 works on YouTube Shorts, TikTok and Instagram Reels.
7. Post with a caption written for each platform, then watch which clips hold viewers and make more like those.

The rest of this guide walks through each step twice: once by hand in any video editor, and once in Clipzing, the AI clipper I build. Use whichever fits how much time you have.

What makes a moment work as a short?

A short is watched by someone who has never seen the long video and may never see it. So the test for a moment is simple: would it make sense to a stranger?

Good candidates usually have three things:

• A clean start. The first sentence does not depend on an earlier one. "And that's why he left" fails; "My co-founder left the week we launched" works.
• One point. A short that tries to carry two ideas usually lands neither. If a passage turns a corner halfway, it is probably two clips.
• A payoff. The clip ends on the answer, the punchline, the number or the turn, not on "so, yeah".

Things that look promising but rarely work: inside jokes, references to "what we said earlier", a slide the viewer can't read at phone size, and long setups that only pay off three minutes later.

How do you turn a long video into shorts by hand?

You can do all of this in any editor that handles vertical sequences and captions: CapCut, Premiere Pro, DaVinci Resolve, Final Cut Pro or similar. The steps are the same.

1. Work from a transcript

Scrubbing a two-hour video to find good moments is slow. Reading is faster. Get a transcript (most editors can transcribe, and YouTube shows one for many videos), read it with a highlighter, and note the timestamps of every passage that passes the stranger test. Mark more candidates than you need.

2. Cut each clip with a little room

Put each passage in its own sequence with a second of handle on either side, then trim the start to the first word that matters. Cut long pauses and false starts, but leave natural rhythm alone; a clip cut too hard sounds robotic.

3. Reframe to vertical

Set the sequence to 1080×1920 (9:16). Scale the landscape footage so the frame is full, then move the crop so the speaker sits in it. In a two-person conversation, the crop has to follow whoever is talking, which means keyframing the position at every change of speaker. This is the most tedious step by hand, and the one where shortcuts show: a crop that snaps back and forth on every short interjection is hard to watch.

If you are not sure of a size, the aspect ratio calculator tells you which platform formats a width and height match.

4. Caption it

Many people scroll with the sound off, so captions carry the clip. Use your editor's auto-captions, then read every line: names, brands and jargon are where transcription goes wrong. Short lines of a few words, timed to the speech, are easier to follow than full sentences. Keep them away from the bottom and right edge, where the apps put their own buttons and text. The safe zone checker overlays the TikTok, Reels and Shorts interface on your frame so you can see what will be covered.

5. Fix the opening

The first seconds decide whether someone keeps watching. If the clip's best line is in the middle, see whether the clip can start there. If it can't, put a short line of on-screen text at the top that says what the clip is about, in plain words. Don't promise something the clip doesn't deliver; viewers who feel tricked leave and don't come back.

6. Export

Export an MP4 (H.264 video, AAC audio) at 1080×1920. One file works on all three platforms. Templates for captions and framing speed this up over time, but finding moments and reframing conversations stay manual work.

How do you do it with Clipzing?

Clipzing does steps 1 to 6 for you and leaves you the decisions. Here is the actual flow in the app.

1. Give it the video

In My Clips, press Create Clip. Choose YouTube link and paste the address, or Upload a file (MP4, MOV, AVI, MKV or WebM, up to 500 MB). Videos can be up to 2 hours long on every plan. Then press Create clips.

If you already know what you want, turn on Direct the clips yourself before you submit. You tell an AI editor what you're after, for example "clip every time I talk about pricing", and then pick which of its suggestions to render. Leave it off and Clipzing picks the best moments itself.

2. Let it find and rank the moments

Clipzing transcribes the video with word-level timing and selects clips from the transcript: passages that start cleanly, make one point and end where a viewer would expect. Candidates that open mid-sentence or depend on earlier context are dropped. Each clip gets a score from 0 to 100 with a short note on what earned it, and arrives with a title, description and hashtags. The score ranks clips from one video against each other; it isn't a forecast of views.

The Starter plan returns up to 5 clips per video. Creator and Agency return as many as pass the checks.

3. Check the framing and captions

Clips are reframed to 9:16 with the frame following whoever is speaking, and the framing is chosen for the kind of video: a two-host podcast, a screen-share lesson and a game stream are each handled differently. Animated captions are burned in with word-level timing.

Open a clip and press Edit clip to change anything in the browser editor: trim to a different sentence by selecting it in the transcript, correct a misheard word, change the caption style, or extend a clip that ended too early. The preview matches the export, and nothing is rendered until you press Export. If you'd rather ask than click, the copilot can retitle clips, fix captions or extend a clip from a chat.

On the Creator and Agency plans you can also add a spoken hook over the opening seconds, or B-roll where it supports what's being said. Neither is on by force; turn them off where the speaker's own opening is stronger.

4. Post or download

From a clip you can Download the vertical file, post it now, or place it on the Calendar. How it goes out depends on the platform (more on that below). Clips are kept for 24, 48 or 72 hours depending on your plan, so download or schedule the ones you want within that window.

5. Automate the next one

If you upload to YouTube on a schedule, Autopilot can watch your channel and clip each new upload by rules you set once: caption style, how many clips, and which connected accounts they go to. Review mode holds every clip for your approval first. Autopilot is on the Creator plan (1 channel) and Agency (5 channels).

How long should each short be?

Let the thought set the length. A clip should be the shortest version that still makes its point and lands its payoff. In practice many good clips fall between 20 and 60 seconds, but that is a result, not a rule: a tight 90-second story will hold people better than a 30-second clip that stops before the answer.

Two quick checks when you're unsure:

• Cover the last ten seconds. If the clip still makes sense without them, cut them.
• Cover the first five. If the clip no longer makes sense without them, they stay, but see whether a better first line exists.

If you're writing a hook or a voiceover to fit a length, the script timer estimates how long a script takes to say.

How should you caption shorts?

Two kinds of text go with a short, and both matter.

On-screen captions are the words burned into the video. Keep them big, short and timed to the speech; highlight the word being spoken if your style allows it. Check spelling of names. Keep them clear of the platform's interface, which differs between TikTok, Reels and Shorts.

The post caption is the text under the video. Each platform shows a different amount before it folds the rest away, so put the point first. The caption length checker counts characters for each platform as you type and shows where Instagram cuts your caption off behind "more". For tags, the hashtag generator suggests a starting set; a few relevant tags beat a wall of generic ones.

How do you write a hook that holds people?

The hook is whatever happens in the first couple of seconds: the first spoken line, the first frame and any text on screen. Some patterns that work without tricking anyone:

• Start on the claim. "Most podcast clips fail for one reason" is a better opening than "So I've been thinking about clips".
• Start on the tension. Open with the question the clip answers, or the moment just before the turn.
• Say what it is. A plain on-screen line ("How we priced our first product") tells the right viewers to stay.

How do you post shorts to each platform?

One 9:16 file works everywhere; what differs is how it gets there.

YouTube Shorts. Vertical videos up to 3 minutes are treated as Shorts. In Clipzing, connect your channel under Accounts, then publish a clip now or place it on the Calendar; it is published as a Short at the time you set. The Starter plan includes 1 connected account and 5 scheduled posts a month; Creator includes 3 accounts and unlimited scheduled posts. See scheduling for the details.

TikTok. Clipzing sends the clip to your TikTok inbox as a draft, straight away. You open it in the TikTok app, add a sound or a final caption if you like, and post it there. It can take several minutes for the draft to appear, and a TikTok draft can't be scheduled for a later time. Autopilot routes clips to TikTok the same way, as drafts.

Instagram Reels. Posting to Instagram from Clipzing isn't available yet. Download the 9:16 clip and upload it in the Instagram app. The cover frame extractor helps you pick a still for the Reels cover, and the feed preview shows how it crops in the profile grid.

When to post. The best time to post tool shows common posting windows per platform in your timezone. Treat them as a starting point, then follow your own analytics once you have a few weeks of posts.

What are the most common mistakes?

• Starting too early. Clips that open on "so", "yeah" or a laugh lose people before the point arrives. Cut to the first line that matters.
• Clips that need context. If a clip only makes sense to people who watched the full video, it won't work for people who didn't.
• Cropping out the person talking. A centre crop of a two-person conversation often frames empty space or the wrong person. Follow the speaker.
• Captions under the buttons. Text that sits where the app draws its like and share buttons can't be read. Check the safe zones.
• Posting everything at once. Ten clips from one episode on one day compete with each other. Spread them out on a calendar.
• Letting clips expire. In Clipzing, clips are stored for 24, 48 or 72 hours depending on plan. Download or schedule the keepers.

Which way should you choose?

Do it by hand if you make a few shorts a month and enjoy the edit. Use an AI clipper when the bottleneck is finding moments and reframing conversations.

If you're comparing tools, our AI clip generator pricing comparison lays out what each one charges, from each vendor's own pricing page. Clipzing's plans are on the pricing page; there is no free plan, and you can cancel anytime.