Import a help centre from a website or files

This feature is in beta: it is still being built and may change.

Bring your existing articles over without copying them by hand. ChaosDesk can read any public help centre, or a ZIP file of Markdown or HTML files.

Every imported article is a draft. Review it and publish it in your help centre when you are happy with it. Running the import again creates nothing new and keeps the edits you made.

From a public website

This works with Zendesk Guide, Freshdesk, Intercom, Help Scout Docs, HubSpot, Document360 and most other help centres and static sites.

1. Open Admin Center, Integrations, Import and pick Help center website.
2. Enter the start address, for example https://help.example.com/hc/en-us. It must start with https://.
3. Optional: set the maximum number of pages (500 by default, up to 5,000) and pages per minute.
4. Run a Dry run, check the report, then run the import.

What ChaosDesk reads

• Only pages on the same site, below the start address. Files are skipped, and so are sign-in, search, request, subscription, vote and comment pages, and anything the site's robots.txt blocks.
• Pages behind a sign-in are not read. Nothing uses your account.
• A page with a heading and enough text becomes an article. Category and section listings do not.
• The category and section come from the breadcrumb trail on each page. Pages without breadcrumbs go to a category named after the site.
• Images in the article are copied into ChaosDesk, also from another host such as a CDN.
• The language comes from the page.

The report warns you when the page limit stopped the import early, and says how many linked pages went unread: raise Maximum pages and run it again, and the articles you already have are kept. It also warns when no article was recognised below the address — start from the help centre's home page in one language.

From a ZIP of files

1. Open Admin Center, Integrations, Import and pick Help center files (Markdown / HTML).
2. Upload a .zip file (at most 20 MB, and 3,000 files counting articles and images together). The page lists the articles and images it found.
3. Run a Dry run or the import.

How the files are read:

• Folders become categories, and sub-folders become sections. Files at the top level go to "Imported articles".
• Markdown files (.md) keep headings, lists, tables, links, code and images. A front matter block can set title, category, section and locale. Otherwise the title is the first heading, or the file name.
• HTML files (.html) are read like a web page.
• Images linked with relative paths are taken from the ZIP and attached to the article.

After the import

1. Open your help centre's articles and filter on drafts.
2. Check the layout and links of each article.
3. Publish what is ready.

Related

• Write an article
• Organise categories and sections
• Moving to ChaosDesk