Rewrite an Article from a URL
What it does: Reads any public article on the web and writes an original piece on the same subject — your angle, your voice, no copied sentences.
Applies to Antradus AI Lite, Publisher & Studio v2.7.0 · Last updated 2026-08-26
When to use it
- A competitor covered a topic and you want your own, better version.
- An industry announcement you need on your site today.
- Turning a source in another language into an article in yours (set your Language — the source language doesn’t matter).
Step by step
- Open the Antradus AI panel on a new post.
- Paste the article’s address into the URL field (
https://example.com/article). - (Optional) Add guidance in the topic box — e.g. "focus on what this means for small businesses". Leave it empty to let the AI pick the angle.
- Choose Style / Tone / Language and click Start Writing.
- The plugin fetches the page, extracts the article text, and writes a fresh piece — plus SEO meta as usual.
What it can and can’t fetch
- ✅ Public articles, blogs, news pages.
- ❌ Paywalled or login-protected pages, and sites that block automated readers (common with academic journals and some news sites). You’ll get a friendly message suggesting an open-access copy — Google Scholar, the author’s own site, or a preprint often works.
- Fetching is security-hardened: the plugin only fetches real public web addresses (internal/private network addresses are refused by design).
You are always told what was actually read
The plugin will not pretend it read a page it couldn’t. Two things can happen:
It refuses outright (nothing is written) when the page isn’t a usable source:
| What we found | What you’ll see |
|---|---|
| A menu, index, category or tag archive — mostly links, little prose | "That page is mostly links… Open the specific article you want and use its own URL." |
| A paywall, login wall, or bot check | "That page is behind a paywall — only the subscribe prompt was readable." |
| A page whose text is built entirely in JavaScript | "There was no readable article text on it." |
| A page that fetched fine but doesn’t actually cover your topic | "Nothing was written, because the page we read is not a usable source: …" — the AI checks the fetched text against your topic before writing, and refuses rather than falling back on its own memory |
It writes, but warns you (a ⚠️ note stays on screen under the buttons) when it read something narrower or wider than you might expect:
- "That page holds roughly 9,300 words — only the first 1,667 were sent to the AI, so anything later on the page is NOT in your article." Long pages are trimmed to keep the request affordable; you’re told exactly how much was skipped.
- "The article area of that page couldn’t be identified, so the whole page was read — including any menus or sidebars."
- "Read only the ‘#section-name’ section of that page — not the whole page." (see below)
Linking to one section of a page
If your URL ends in an anchor — https://example.com/docs/#installing-the-plugin — the plugin reads only that section, not the whole page, and tells you it did. This matters on documentation hubs, FAQ pages, and any long page that holds many topics at once: without it you’d get the site’s table of contents instead of the part you meant.
If that anchor doesn’t exist on the page, the whole page is read and you’re warned so you can check the result.
Tips & gotchas
- The output is original writing about the same subject — not a paraphrase. Facts carry over; sentences don’t.
- Combine with 🌐 Search the web if the source is older and you want the AI to check for newer developments.
- Want two sources blended into one? That’s Merge Two Sources.