Why most repurposing reads as spam
The default repurposing workflow — auto-clip 60-second snippets of a long-form video, cross-post to every platform with the same caption — produces content that feels like obvious leftovers. Audiences smell it. Engagement collapses. The creator concludes "repurposing does not work" and goes back to making 7 pieces of content per week from scratch.
What actually does not work is naive repurposing. Good repurposing treats each platform as a distinct audience with distinct norms. The pillar piece (long-form video, podcast, newsletter) is the source. The 10+ derivatives are each authored — even if AI does the heavy lifting — to fit the platform they live on.
Done well, the same pillar piece reaches 5-20× the audience it would have on its native platform alone. Done badly, it reaches less audience than the pillar alone because algorithms suppress obvious cross-posts.
The pillar → derivative tree
One pillar piece (say a 25-minute YouTube video) maps to roughly: 1 YouTube long-form upload (the pillar itself), 3-5 short-form clips for IG Reels / TikTok / YouTube Shorts (each 30-60 seconds, each with its own native hook), 1 LinkedIn carousel (~7-10 slides, text-heavy, no video), 1 X/Twitter thread (5-10 tweets), 1 newsletter section that summarizes and links back, 1 community/Discord post for your inner audience.
That is 8-10 derivatives from one pillar. The wrong way to do this is "edit the same 60-second clip 8 times." The right way is to extract 8 different angles from the source — the punchy quote becomes a tweet, the framework becomes a carousel, the demo becomes a clip, the takeaway becomes a tip.
Each derivative needs its own hook tuned for the platform. A LinkedIn carousel hook is different from a TikTok hook is different from a Twitter thread hook. AI can generate all of these from the same source, but you have to prompt for the right format per derivative, not "make me 10 clips."
AI prompt for 10 derivatives from one pillar
- Role: You are a content repurposing strategist for a solo creator.
- Input: Here is the full transcript of my [25-min YouTube video / podcast / newsletter]: [paste]
- Produce: 1) Three 45-60 second short-form clip ideas with timestamps, hooks, and platform (IG/TikTok/Shorts).
- 2) One LinkedIn carousel outline (slide-by-slide, 7-10 slides).
- 3) One X/Twitter thread (5-10 tweets, including the opening hook and the closing CTA).
- 4) One newsletter section (~150 words, with a link back to the pillar).
- 5) One community/Discord-style post (casual, ~80 words).
- Constraint: Each derivative must have a platform-native hook. Do not reuse the YouTube title as the hook anywhere else.
The 2026 tool stack
Opus Clip / Vizard / Submagic ($30-100/mo). Auto-extract short-form clips from long-form video with AI-generated captions and thumbnails. Best fit when you record long-form video weekly and need 3-5 clips per episode. Quality varies — review every clip before posting.
Castmagic / Capsho ($24-50/mo). Podcast-first transcription + blog/social repurposing. Generates social posts, blog draft, email section from one audio upload. Best fit for podcasters.
Claude or GPT-4 via Projects/GPTs ($20/mo). The flexible center. Custom prompts produce LinkedIn carousels, X threads, newsletter sections from a transcript. Best when you want control over voice across derivatives.
Buffer / Hypefury / Typefully ($15-30/mo). Multi-platform scheduling. Post the derivatives to their respective platforms with native scheduling so each one drops at its own ideal time, not all at once.
Total stack: ~$80-150/month. Replaces 6-10 hours/week of manual repurposing work for active creators. The math works at any meaningful audience size.
The "no slop" rules
Rule 1: Never auto-publish AI-generated derivatives. Every one needs a 30-second human review for voice, accuracy, and platform fit. The whole stack collapses if you skip this — audiences detect zero-touch slop within 2-3 posts.
Rule 2: Stagger publication across days, not minutes. The same pillar dropping as 8 derivatives in one afternoon screams "automated." Spread across 5-10 days; each platform gets the post when its audience is most active.
Rule 3: Each derivative should make sense to someone who has never seen the pillar. The Twitter thread should not say "watch the full video for context" — it should deliver value standalone, with a link to the pillar as a "go deeper" option.
The 30-second review test
Before publishing any AI-generated derivative, read it aloud once. If you would not have written it that way, edit until you would have. The 30-second pass per derivative is what separates "AI helps me ship more" from "AI made me sound like a generic bot."