ShortsCutter

Guide

How to Cut a Long Video into Shorts, Step by Step

Your long videos already contain your best Shorts — the problem is the hours it takes to find, reframe and caption them by hand. This guide walks through the full manual workflow for turning a long video into clips that are ready to post, then shows how ShortsCutter automates the same steps.

It applies to any long-form source: a talk, a podcast episode, an interview, a tutorial or a livestream recording. If you make speech-driven videos and want to batch-create Shorts from them, this is the whole process.

Part 1 — The manual workflow

Six steps, in order. Skipping one usually shows up later as a clip that gets scrolled past.

STEP 1

Find the moments worth cutting

Scrub through the video (a transcript makes this much faster) and mark every moment that could stand alone. The standards that separate a clip from a random excerpt:

It opens on a hook — a question, a surprising claim, a specific number, or the start of a story.
It carries emotion — excitement, disbelief, laughter, a strong opinion. Flat, informational stretches rarely travel.
It is self-contained — a viewer who has never seen the long video understands it cold.
It ends cleanly — on a complete thought, ideally a line worth quoting or retelling.

Mark more candidates than you need. Choosing the top three from ten candidates beats polishing the first moment you stumble on.

STEP 2

Set the in and out points

Trim each candidate to a single complete idea. For talks and podcasts, 30–60 seconds is the practical range: long enough to land the point, short enough to hold a cold audience. Start and end on complete sentences — a clip that begins mid-word or cuts off the punchline feels broken even if the content is strong. If a moment needs a long setup, either find a later starting point or split it into two clips.

STEP 3

Reframe the picture to vertical 9:16

Landscape footage has to become vertical without losing the speaker. You have two honest options:

Center crop — zoom into the middle of the frame. Sharp, but on wide shots it can cut off a second speaker or a slide at the edges.
Blur-pad (pillarbox) — keep the full frame, scale it to the width of the vertical canvas and fill the space above and below with a blurred copy of the video. Nothing is lost, which is why it is the safer default for talks and interviews.

Whichever you choose, check the result on a phone: faces and any on-screen text must stay clear of the very top and bottom of the frame, where the apps draw their own titles and buttons.

STEP 4

Add burned-in captions

A large share of short-form viewing happens muted, so captions are not an accessibility extra — they are the product. Style rules that work:

Burn them in (rendered into the video file) so they look identical on YouTube Shorts, TikTok and Instagram Reels.
Big and brief — a few words at a time, in short phrases matched to the speech, rather than full sentences in small type.
Bottom center, inside the safe zone — high enough to clear the description text and buttons each platform overlays at the bottom and right edge of the screen.
High contrast — white text with a dark outline reads over any background.

STEP 5

Fix the first 2 seconds

A Short is decided almost immediately. The opening of your clip is whatever the speaker happens to say first, so make it work: start on the question, the claim or the story — not on “so, um, in this video…”. If the best line sits a few seconds into your marked segment, move the in-point to it. Skip intros, logos and greetings entirely; the long video's branding does nothing for a viewer meeting you for the first time.

STEP 6

Export to each platform's specs

One master export covers most placements, with alternates when a feed crops differently:

Placement	Aspect ratio	Resolution	Notes
YouTube Shorts	9:16	1080×1920	Up to 3 minutes; repurposed highlights usually land at 30–60 seconds.
TikTok	9:16	1080×1920	Keep captions clear of the caption text and action buttons overlaid at the bottom and right.
Instagram Reels	9:16	1080×1920	Feed previews crop toward 4:5 — keep the subject and captions centered.
Feed / square placements	1:1 or 4:5	1080×1080 / 1080×1350	Useful where a full-height video is letterboxed or cropped.
Landscape reuse	16:9	1920×1080	For clips embedded back into essays, presentations or compilations.

Part 2 — The same workflow, automated with ShortsCutter

ShortsCutter is the long-video-to-Shorts maker built around exactly the steps above. You send a YouTube link or your own video file; it returns post-ready clips. Here is how each manual step maps:

Finding moments (Steps 1–2). The video is transcribed with word-level timestamps, then every candidate window is scored on the same standards you would apply by hand: opening hooks (question words, numbers, story markers, direct address), emotion-word density, and pacing — plus two things humans check instinctively: whether the window starts and ends on complete sentences, and how confident the transcription is. The top-scoring, non-overlapping windows become clips. It is explainable heuristics, not an LLM: every clip comes with its score breakdown, so you can see why it was picked.
Reframing (Step 3). Clips are rendered to 9:16 by default using the blur-pad approach, so nothing at the edges of the original frame is lost; a center-crop mode is available when you prefer a full-bleed picture.
Captions (Step 4). Captions are generated from the transcript, split into short phrases, and burned in — large white text with a dark outline, bottom center.
Sizes (Step 6). Beyond 9:16, the same clip can be rendered as 1:1, 4:5 or 16:9 — or any custom width×height — with caption size and margins automatically scaled to the canvas.
Batch work. Each long video yields multiple clips in one run, which is what makes a weekly podcast or talk series practical to repurpose: send the episode, get the Shorts.

One honest limit, stated on the product page too: the method reads the words. It works best on talks, podcasts, interviews and tutorials. Content carried by visuals, music or gameplay gives the scoring too little to judge, and no face-tracking crop is applied to wide multi-person shots — the blur-pad default exists precisely so nobody gets cropped out.

Manual vs. automatic: side by side

Step	By hand	With ShortsCutter
Find highlights	Re-watch or scrub the whole video; judge candidates from memory	Full transcript scored on hooks, emotion and pacing; top windows picked automatically
Know why a clip was picked	Your own notes, if you wrote any	A written score breakdown for every clip
Set in/out points	Manual trimming, listening for sentence boundaries	Windows anchored to the transcript; scored partly on starting and ending on complete sentences
Reframe to vertical	Re-compose each clip in an editor	Automatic 9:16 blur-pad (or center crop), plus 1:1, 4:5, 16:9 and custom sizes
Captions	Type or auto-caption per clip, then restyle per platform	Burned in from the transcript, styled once, identical on every platform
Batch a series	Repeat the whole process per episode, per clip	Multiple clips per long video in one run
Creative control	Total — every decision is yours	You judge the finished clips and the reasons behind them; best when you want the first pass done for you

FAQ

How long should a Short cut from a long video be?

For speech-driven content, 30 to 60 seconds is the practical sweet spot: long enough for one complete idea, short enough to hold attention. YouTube Shorts allows up to 3 minutes, but a repurposed highlight rarely needs it — if the moment needs more than a minute of setup, it is probably two clips, not one.

What makes a moment in a long video worth turning into a Short?

The strongest candidates share four traits: they open with a hook (a question, a surprising claim, a number or a story starting), they carry emotion, they make sense on their own without the rest of the video, and they start and end on complete sentences. A moment that needs the previous five minutes to make sense will not survive as a standalone clip.

Can I batch-create Shorts from a podcast or talk?

Yes — that is the normal way to work. One long video usually contains several clip-worthy moments, so the workflow is: transcribe the episode, score every candidate window, and cut the top few as separate Shorts instead of hunting for a single perfect moment. ShortsCutter does this per video automatically, and the founding beta covers up to 30 long videos per month with unlimited clips.

Do Shorts cut from long videos need captions?

Yes. A large share of short-form viewing happens with the sound off or in noisy places, and captions also make the hook readable in the first two seconds. Burned-in captions — large text, a few words at a time, placed at the bottom center but clear of the platform's own overlays — travel with the file to every platform, unlike platform-specific caption features.

What size should I export: 9:16, 1:1, 4:5 or 16:9?

9:16 (1080×1920) is the default for YouTube Shorts, TikTok and Instagram Reels because it fills the phone screen. 1:1 and 4:5 suit feed placements where a full-height video gets cropped or letterboxed, and 16:9 is useful when the clip goes back into a landscape context such as a video essay or a presentation. ShortsCutter renders all of these presets, or any custom width×height, with caption size and margins scaled to the canvas.

How does ShortsCutter pick the moments without an AI black box?

It transcribes the video, then scores every candidate window with explainable heuristics: opening hooks (question words, numbers, story markers, direct address), emotion-word density, pacing (speech rate in a natural band), whether the window starts and ends on complete sentences, and transcription confidence. Every clip ships with its score breakdown, so you can see exactly why it was picked — no LLM and no opaque ranking.

What kinds of long videos work best for this?

Speech-driven formats: talks, podcasts, interviews, tutorials, lectures and livestream commentary — content where the value is in what is said, so a transcript reveals where the highlights are. It is a poor fit for content carried by visuals, music or gameplay, where the transcript holds too little information to judge a moment.

$19 / month — founding beta

Send a long video, get ready-to-post Shorts with burned-in captions and a score breakdown for every clip. Founding members keep the $19 price.

Subscribe — $19/month Apply for a free trial cut first

See real, unedited output on the ShortsCutter product page.