How to Make YouTube Shorts From a Long Video

YouTube Shorts can now run up to three minutes long, but the best-performing ones still sit closer to 20 to 60 seconds. The real skill in turning a long video into Shorts is not trimming, it is finding which few minutes out of an hour are a complete story on their own. Here is how to find them and cut them without losing the source video's sync.

The spec

What actually counts as a Short in 2026

YouTube extended the Shorts length limit from 60 seconds to 3 minutes in October 2024, and that ceiling is still current. Licensed music clips inside a Short are still capped at 60 seconds regardless of the video length, which matters if you plan to score a clip with a commercial track. In practice, the platform's own data keeps rewarding shorter cuts: the best-performing range across large creator datasets sits between 20 and 60 seconds, with a soft ceiling around 90 seconds before completion rate starts to drop. Treat 3 minutes as the legal limit, not the target.

The same clip usually travels well to Instagram Reels and TikTok with only minor adjustments. Reels tops out at 3 minutes too, and TikTok allows uploads as long as 60 minutes though its own short-form culture still rewards clips well under a minute. All three platforms favor the same 9:16 vertical frame and the same sound-off-by-default viewing pattern, so a clip cut and captioned once for YouTube Shorts generally holds up on the other two with a re-export and a platform-appropriate caption, not a full rebuild.

Find the moment

Read the transcript, don't scrub the footage

Scrubbing a 60-minute video looking for clip-worthy moments means watching most of it at close to real speed, multiple times, per clip. Reading the transcript lets you scan for a complete idea, a strong opinion, a story with a beginning and an end, far faster than watching does. Mark candidate passages on the text first, then go confirm the delivery on video only for the ones that survive the read.

Media reporter Max Tani made this point about legacy TV in a Semafor interview with the Colin and Samir Show: formats like Saturday Night Live and The Daily Show now generate significant YouTube engagement specifically because they are "made to be clipped," meaning individual segments already function as complete, self-contained moments. That is the bar for a Short pulled from a podcast or interview too: if the segment does not stand alone without the surrounding hour, it is not ready to cut yet.

Creators building large audiences on YouTube treat this as a discovery layer now rather than an afterthought. Podcasters like Lex Fridman and the team behind The Diary of a CEO regularly turn single long-form conversations into a dozen or more standalone clips, each competing for its own view on its own merits, not as a preview of the full episode. The long-form video and the Shorts pulled from it are functioning as two separate products built from one recording, and the transcript is what makes producing that volume of clips realistic on a normal schedule.

Lead with the hook

Win the first two seconds or lose the scroll

A Short has to earn attention before the viewer has committed to anything. Open on the most interesting frame or line, not a wind-up. If the best line in the segment is the third sentence, cut the first two and start there; you can always let context fill in as the clip plays rather than front-loading a setup nobody asked for.

Reframe to vertical

Turn 16:9 into 9:16 without losing the point

Source footage shot for a landscape frame needs to be reframed, not just cropped down the middle. Center the crop on whoever is speaking, and if the framing needs to move between two people trading a conversation, cut on the switch rather than panning, since a pan reads as unintentional camera movement in a clip that otherwise looks locked off. Letterboxing the original 16:9 frame inside a vertical canvas is the fallback when reframing genuinely loses something essential, like an on-screen graphic, but it should be the exception, not the default.

Two-person conversations are the trickiest reframe. Panning or bouncing the crop back and forth on every line reads as nervous camera work, even though nothing in the original shot actually moved. A cleaner approach snaps to whoever is speaking at natural sentence or turn boundaries, holding the frame steady in between, so the reframe looks like an editorial choice rather than an automated camera trying to follow a conversation in real time.

Tighten the talk

Remove fillers and dead air from the selected clip

Once you have a candidate passage, tighten it the same way you would a full talking-head edit: cut filler words, false starts, and any pause that is not doing rhetorical work. A Short has no runway for a slow start or a mid-sentence stumble, so this pass should be more aggressive here than it would be on a long-form version of the same content.

Watch the runtime as you cut, not just the content. A 45-second clip with a clean hook and a tight middle almost always outperforms a 90-second version of the same idea padded with a slower setup, even though the longer cut technically contains more information. If tightening the audio takes a clip from 70 seconds down to 40 without losing the point, that is a better outcome than preserving every word of the original delivery.

Caption everything

Assume the sound is off

Shorts and Reels are consumed on autoplay with the sound muted by default for a large share of viewers. Burned-in captions are not optional polish, they are the difference between a clip that communicates and one that gets scrolled past silently. Keep captions synced tightly to speech and legible against moving footage, not just technically present.

Worked example

One hour-long podcast, six Shorts

A one-hour interview typically yields somewhere between six and ten usable Shorts once you have read the transcript for complete, self-contained moments. Not every strong quote works; a great line that depends on ten minutes of prior context is a long-form highlight, not a Short. Reading first means you can triage this quickly: mark every candidate passage, then reread each one asking whether a viewer with zero context would understand and feel something from it alone. Cut only the ones that pass, reframe to vertical, tighten the audio, caption, publish.

Spread across a week, six to ten Shorts from one recording session gives a channel a full posting cadence without a second shoot. That is the real economic case for this workflow: the cost of finding and cutting the clips is mostly the read-through, which happens once, while the source footage itself was already paid for by the long-form production. Treating a single interview or podcast episode as a bank of future short-form content, rather than a one-time upload, is what turns a normal recording schedule into a much larger publishing footprint without proportionally more shoot days.

Common mistakes

Where Shorts from long-form video go wrong

Publishing a clip that is really a trailer for the full episode, ending on a cliffhanger with a "watch the full video to find out" energy, is the most common failure. Shorts perform on completion and rewatch, and a clip that deliberately withholds its own payoff trains the algorithm to stop showing it. The second mistake is over-relying on one long clip when three shorter, tighter ones from the same segment would each have a cleaner hook. The third is skipping the reframe entirely and uploading a letterboxed 16:9 video with black bars top and bottom, which reads as low effort next to a properly reframed clip in the same feed.

A less obvious mistake is treating every clip's caption pack and hook the same way regardless of platform. The line that works as a punchy on-screen hook for a YouTube Shorts thumbnail moment does not automatically make a good TikTok caption or Instagram Reels cover frame, since each platform's discovery surface presents that first frame slightly differently. Reviewing the hook frame per platform before publishing takes a minute and meaningfully changes how many people stop scrolling.

How ScriptCut fits

Where a transcript-based tool helps in this workflow

ScriptCut's AI Clips feature scores a long transcript for self-contained, clip-worthy passages and surfaces them for review, which speeds up the read-and-mark step above without replacing your judgment on what actually earns a publish. Selected clips carry word-level timecode into your export, so the reframe and trim work is starting from an already-precise in and out point rather than a rough guess.

None of this replaces a human decision on what actually gets published. AI scoring is useful for triage across a long recording where reading everything closely is not realistic, but the final call on whether a clip is genuinely self-contained, funny, or useful still belongs to whoever knows the audience. Treat the tool as a way to narrow sixty minutes down to twelve candidates faster, not as a publish button.

Sources

• Semafor: Mixed Signals, how YouTube became TV, with Colin and Samir
• YouTube Help: Create Shorts on YouTube
• Podcast Clips: How I Turn Long-Form Videos Into Short Clips, YouTube

Related reading: How to edit a talking-head video, What is a paper edit?, What is text-based editing?, Otter.ai alternative for video editing, Veed alternative for professional editors.