What it does
Content Repurposing Agent solves the tax every content team pays: one good piece gets published, and then someone spends the rest of the day reshaping it by hand for five channels that each want something different. You paste one source — a blog post, a webinar transcript, a report — and it returns the package. A LinkedIn post with a hook, key-point bullets and a call to action. An X thread with a hook tweet, one tweet per key point and a closing CTA. A newsletter section with a headline, intro, takeaways and link. A sixty-second video script laid out beat by beat. And carousel slide copy, one slide per key point. The counts are the part worth trusting. Tweet count, slide count, character count, script seconds — all computed deterministically from the key points the analyzer extracted, not asserted by a model that thinks eight tweets sounds about right. A LinkedIn post that reports 512 characters is 512 characters, and each output carries a flag for whether it is within the channel's limit. The language model writes only a short editorial note — which channel to lead with and why — and is explicitly barred from changing a count or inventing copy. Every channel is its own tool, which is how this actually gets used. Most days you want one thing: `write_x_thread` for the thread, `write_carousel` for the deck, `write_newsletter` for this week's issue. `extract_key_points` returns the extraction alone, which is useful when you want to check what the agent thinks the piece is about before it writes anything. `run_full` produces all five. You can also pick a subset of channels and pass a brand tone that is applied per channel. No connectors and no credentials. The agent returns copy for you to schedule; it does not post, and it does not reach into a CMS. Built for content leads and social managers.
Example prompts
- Adapt this blog post for LinkedIn, X and our newsletter
- Turn this webinar transcript into a carousel
- Write just the X thread from this piece
