If you already have documentation on a website — your own blog, a Notion doc, a Confluence page, a Gitbook — you can pull it into HelpGuides with a single URL. The importer fetches the page, removes navigation and layout chrome, and converts the content to clean markdown in the editor.
This feature requires no AI credits and is available on all plans.
How to use it
Step 1 — Open the URL import modal
Go to Editor → Articles → New Article. In the AI Assist bar at the top of the editor, click From URL.
The same option is available when editing an existing article — useful for bringing in a reference page to build on.
Step 2 — Paste the URL and import
Paste the full URL of the page you want to import (must start with https:// or http://) and click Import. Press Enter to import without clicking.
The importer fetches the page server-side. This typically takes a few seconds. If the URL is unreachable or times out after 15 seconds, you'll see an error.
Step 3 — Review what was imported
Once the import completes, the editor is populated with:
| Field | Source |
|---|---|
| Content | The page body, converted from HTML to markdown |
| Title | og:title → <title> tag → first <h1>, in that order |
| Description | og:description → <meta name="description"> |
| Slug | Auto-derived from the imported title |
| Meta title / description | Copied from title / description |
These fields are only filled in if they were previously empty — importing into an article that already has a title won't overwrite it.
Images in the imported content are preserved as links pointing to the original source URLs. They are not downloaded or re-hosted.
Step 4 — Edit and publish
The imported content is a starting point, not a finished article. Review the markdown, clean up anything that didn't convert cleanly, assign a category, and publish when ready.
What the importer strips
The server strips the following before converting to markdown:
- Navigation bars (
<nav>) - Headers and footers (
<header>,<footer>) - Sidebars (
<aside>) - Scripts, styles, and forms
- HTML comments
It tries to identify the main content area by looking for a <main>, <article>, or common content container — falling back to the full <body> if none are found. On some pages, residual links from menus or footers may still appear in the imported markdown.
Limitations
The page must be publicly accessible. The importer fetches the URL as an anonymous visitor. Pages behind a login, paywall, VPN, or IP allowlist will return an error.
JavaScript-rendered content is not supported. The importer reads raw HTML. If the page relies on JavaScript to render its content (single-page apps, client-side documentation generators), the import may return an empty or partial result.
Complex layouts convert approximately. Tables, multi-column layouts, accordions, and interactive components are simplified or dropped during HTML-to-markdown conversion. The more straightforward the source page's HTML, the cleaner the result.
15-second timeout. Slow or unreliable servers may cause the import to fail. Try again, or copy-paste the content manually if the URL consistently times out.
Best practices
Only import content you have the right to use. This is the most important rule. Use this feature for your own existing documentation, internal wikis you're migrating, or content you have explicit permission to republish. Do not import third-party articles without permission.
Prefer simple article pages. Blog posts, plain docs pages, and help center articles convert well. Marketing landing pages, dashboards, and documentation generated by JavaScript-heavy tools tend to produce noisy results.
Always edit before publishing. Imported content uses the source page's structure and wording. Rewrite it to match your portal's tone, remove any references to the original site, and add any context specific to your product.
Update the title and slug. The imported title comes from the source page's SEO metadata, which may be worded for a different audience. Update both to fit your portal's conventions.
Check and replace images. Imported images point to the original source's CDN. If that site goes down or the image URL changes, your article will show broken images. Re-upload critical images to your own portal for reliability.
Use it for a first draft, not a finished article. The importer saves you the work of reformatting — it doesn't save you the work of writing. Treat the result the same way you'd treat any generated content: edit, verify, and own it before you publish.
Last updated: June 6, 2026