How to repurpose long-form video into shorts (without burning out)

Whatever workflow you pick, you're doing the same five things:

1. Find the moments. Six to ten clip-worthy moments per hour of source footage is typical.
2. Trim with hooks intact. The first second has to make someone stop scrolling. Cutting too early or too late kills the clip.
3. Reframe to vertical. Static center-crop loses the speaker. Face tracking keeps them in the middle of a 9:16 frame.
4. Add word-highlighted captions. 80% of mobile viewers watch with sound off. Captions are non-negotiable.
5. Post on cadence.One short a day per channel beats a burst of six and silence for a week. Schedule, don't batch-publish.

The three workflows below differ in which steps you do by hand, and how much you spend per clip in either time or money.

Watch the source video, drop markers on clip-worthy moments, export each segment, vertical-crop in a 9:16 timeline, add captions in CapCut or by hand, post each one with a hand-written title.

This is what most creators start with, and it's also what most quit after a month. The clips look great, but the time cost scales linearly with output — if you want one short a day and each takes 25 minutes, that's three hours a week before you've thought about thumbnails or scheduling. Solo creators with strong editing chops sometimes stick with this for the first 90 days; nobody we know runs it past that.

Use a tool like Descript, CapCut, or Opus Clip to handle transcription, vertical reframe, and captions. You still pick the moments yourself and write the metadata.

This is where most creators land after the manual phase. Captions and reframing get cheap and fast. The two remaining time sinks are watching the source to find moments (you're still doing this), and writing titles + hashtags for each clip. Output ceiling is roughly two-to-three clips per hour of work — fine for a solo channel, painful if you're running multiple.

Hand a long-form URL to a tool that transcribes, ranks every segment for hook strength, reframes, captions, and posts on a schedule. ReelBatch is one; OpusClip and Vizard are others. You review the picks and approve a queue.

The ranker is the make-or-break component. A bad one returns six mid moments and you're back to manual scrubbing to find anything publishable. A good one returns picks that overlap with what the creator's own editor would have chosen — the bar we benchmark against. AI quality matters more than UI: you can forgive a clunky dashboard, you can't forgive bad clips.

• Sub-1k subscribers, learning your voice: fully manual for the first 30–60 clips. You'll absorb what makes a hook land. Skip ahead and you won't know why your AI tool's picks are bad when they are.
• Solo creator publishing weekly: partial automation. Pick your own moments, let tools handle the mechanical work. Adds 90 minutes a week max.
• Multiple channels or daily output: full automation. The math stops working otherwise — manually clipping 30 shorts a week against a day job is a recipe for quitting.

1. Center-cropping instead of face-tracking. Static crop puts the speaker off-screen the moment they shift in their chair. Face-tracked crop is table stakes.
2. Captions without word-by-word highlighting. Plain subtitles work; word-highlighted captions outperform them by 20-30% on retention in our internal tests because they pace the read.
3. Posting six clips on Monday and nothing the rest of the week. The algorithm rewards consistent presence. A scheduler that spaces clips out beats batch publishing every time.
4. Re-using the same hook across channels. If you run multiple accounts, vary the title and first-line hook per channel — the platforms detect duplicate content and suppress the duplicates.
5. Skipping the hook.A clip that opens with “so anyway, what I was saying” is dead. The first 0.5 seconds need a question, a stat, or a specific claim.

We built ReelBatch for the workflow-3 cohort: creators with steady long-form output who want full automation without sacrificing pick quality. The viral ranker uses Claude to score every segment for hook strength, emotional peaks, and standalone watchability — the picks have averaged ~60-70% overlapwith what the creator's own editor chose on the test videos we benchmarked against.

Vertical reframing uses face tracking by default, captions are ASS-burned with word-by-word highlighting, and posting goes to YouTube Shorts on a schedule you set. TikTok and Reels are rolling out as platform partner approvals land.

No card for the sample request · paid plans include a 7-day trial

• YouTube Shorts automation — how the posting and scheduling pipeline works.
• Faceless YouTube clip operator workflow — for creators running multiple channels.