← Quirewell
DocumentationPrivacyEULASupport

Quirewell: Help Center Article Importer

Quirewell imports Word, Markdown, HTML, Google Docs and Confluence exports and CSV files into your Zendesk help center as articles, with their pictures, links, translations and structure. It runs in your browser inside Zendesk, shows a preview of every article first, and can undo an import.

Getting started

  1. Install Quirewell from the Zendesk Marketplace (Standard plan, $99 a month per Zendesk account, 14-day trial).
  2. Open Quirewell from the left navigation bar in Zendesk.
  3. Pick files, a folder or a zip, or drop them on the page.
  4. Choose where files go and the article settings, click Preview import, check the rows, then click Import.

Run imports as a Zendesk admin so the import history is saved and undo stays available after the page is closed. Agents who can manage the knowledge base can import too; their imports can be undone only while the page stays open.

What you can import

SourceHow to get the fileWhat carries over
Microsoft WordSave as .docx (Word 2007 or later). Old .doc files: open and save as .docxTitle (the Title style or document properties), headings (heading styles or outline levels), bold, italic, underline, strikethrough, superscript, subscript, inline code (code styles or monospace fonts), code blocks, quotes, nested bulleted and numbered lists, tables with header rows and merged cells, links, bookmarks, pictures with alt text, footnotes
Google DocsFile > Download > Web page (.html, zipped)Title, headings, bold, italic, underline, monospace, lists, tables, links (Google's redirect links are unwrapped), pictures from the zip. Comments are removed
ConfluenceSpace export as HTML (zip)Page title, content, code blocks with their language, info and warning panels (as quotes), pictures and attachments shown on the page, links between exported pages. Breadcrumbs, the attachment list and page chrome are removed. A page's first breadcrumb below the space home becomes its section
Markdown (.md, .markdown, .txt)Any editor or docs site sourceGitHub-style tables, fenced code with its language, task lists (as plain list items), strikethrough, links, pictures, raw HTML; optional front matter (below)
HTML (.html, .htm)Any web page or exportHeadings, text formatting, lists, tables, code, quotes, links, pictures, YouTube, Vimeo, Wistia and Loom players
CSV (.csv, .tsv)One row per articleTitle and body columns (body as HTML or Markdown), optional section, section_id, category, labels, locale, source_id, draft, promoted, format

Pictures: PNG, JPG and GIF are used as they are; WebP, BMP, SVG and AVIF are converted to PNG in your browser. Each picture can be up to 20 MB. Pictures from the web (an http address) stay where they are. One import can hold up to 500 MB of files.

Left out, and named in the preview: Word text boxes, charts, shapes, equations (kept as plain text), comments, endnotes, page headers and footers and the table of contents; scripts, styles, forms and other embedded frames in HTML; PDFs, .doc, .odt, .rtf and Pages files.

Where articles go

  • Folders become sections: a file in Billing/Refunds/ goes to the section "Refunds" in the category "Billing". A file in one folder (Refunds/) goes to that section in the category you picked. Files at the top go to the category and section you picked. Folders deeper than two levels are joined into one section name ("Refunds / Older").
  • Everything into one section: every file goes to the section you picked.
  • A folder that every file shares (the folder you zipped) is skipped.
  • Categories and sections are matched by name, ignoring case. Missing ones are created, in your default language, and marked "new" in the preview.
  • Front matter or CSV columns override the folders for that file.

Markdown front matter

---
title: Resetting your password
category: Account
section: Sign-in
labels: [password, login]
locale: fr
draft: true
promoted: false
---

CSV columns

title and body are required. Optional: format (markdown or html, otherwise detected), section, section_id (an existing section's id), category, labels (comma or semicolon separated), locale, source_id (rows with the same source_id in different languages become one article with translations), draft, promoted. Header names are matched loosely ("Article Title", "label_names"), and comma, semicolon and tab separated files work.

Translations

Files that are the same document in different languages become one article with translations:

  • name them with the language code before the extension: refunds.docx, refunds.fr.docx, refunds.de.docx, or
  • put each language in a top folder named after it: en-us/Billing/refunds.md, fr/Billing/refunds.md, or
  • give CSV rows the same source_id and different locale values.

The language you pick in Quirewell is the article's source language. A document with no file in that language is created in its own language. Languages must be turned on in your help center.

Duplicates

The preview marks a row as a duplicate, and leaves it unticked, when:

  • an article with the same title (ignoring case and punctuation) is already in the target section, or
  • the same file was imported before in an import that was not undone.

An article with the same title in another section is only a note. Tick a duplicate to import it anyway.

Article settings

Visible to (everyone or a user segment), permission group (who can edit), labels added to every article (plus any from front matter or CSV), drafts (on by default), promoted, and comments off. Subscribers are not notified about imported articles.

Undo

History lists every import with its name, time, who ran it and what it created. Undo:

  1. archives the import's articles (restore them from the Guide archive if needed),
  2. deletes the sections it created when they hold no other articles,
  3. deletes the categories it created when they hold no sections,
  4. deletes the pictures it uploaded.

Edits made to the imported articles after the import are archived with them.

Troubleshooting

  • "Image not found in the files you picked": the document points at a picture that is not in your selection. Pick the folder or zip that holds it.
  • A picture was left out with "could not send the image to Zendesk file storage": the browser could not upload it. Try again; if it repeats, email us with the browser name.
  • Permission error: the agent needs to manage the knowledge base and be a member of the chosen permission group. Publishing directly (drafts off) also needs publishing rights.
  • Translation skipped: turn the language on in Guide settings.
  • Undo is missing after a reload: the import was run by an agent who is not an admin.
  • Imports go to the wrong help center: Quirewell imports into the help center of your default brand.

Your data

Files are read in your browser and go only to your Zendesk account. Quirewell has no server, keeps no copies, has no export, and uses no AI. The import history is kept in Quirewell's own app setting in your Zendesk and holds ids only. Privacy policy: https://greatwork.company/apps/quirewell/privacy. Terms: https://greatwork.company/apps/quirewell/eula.

Support

Email hello@greatwork.company. We reply within one business day.