Importing Your Content from Another Service

Ready to bring your existing help content into HelpDocs? You can import from another help desk, from a repository, from a spreadsheet, from a folder of documents, or straight off your current website. Head to Settings > Migrations to start 🎉

How an import runs

1. Pick a source and give it what it needs: a credential, a file, or a URL.
2. We prepare it. We fetch your content, convert it, and stage it. Nothing has been written yet. Large sites and big uploads can take a few minutes, and you can close the tab and come back.
3. You review it. You'll see how many folders and articles we found, a list of the article titles, and any notes about content that didn't convert cleanly. This is where you map CSV columns, match authors to your team, and confirm the mode and destination folder.
4. You press Import. We write the content, reindex search, set up redirects where the source supports them, and copy any images across to HelpDocs so they don't break when your old site goes away.

An import in progress shows its status in the Migrations tab. If you refresh the page or come back later, the wizard reopens whatever job you had running.

Status	What it means
Preparing	We're fetching and reading your content.
Ready to review	Parsed and waiting for you. Nothing has been written.
Importing	Writing content into your knowledge base.
Completed	Finished.
Failed	Something went wrong. Nothing further was written.
Cancelled	You stopped it. Anything already written stays and is searchable.

Choosing an import mode

Every import asks how it should treat the content you already have.

Mode	What it does	Use it when
Add alongside default	Creates new folders and articles. Nothing existing is touched or deleted.	Almost always. This is the safe one.
Update existing	Matches each incoming article against one you already have, by external ID or by HelpDocs article ID, and updates it in place. Anything with no match is created.	You're re-running an import to pick up changes made on the other side.
Replace everything	Deletes every article, folder and redirect in the account, then imports. You have to type REPLACE to confirm.	Starting again from scratch on an account with nothing you want to keep.

Importing into a folder

Every import except Replace everything can drop its content inside an existing folder, using Import into folder. Any folders the import creates become subfolders of the one you pick. It's the tidiest way to bring in a second source without it landing all over your top level, and it makes the import easy to find afterwards.

Where you can import from

From another platform

We read your content over the platform's own API. Nothing is changed or removed on the other side, and you can run these as often as you like.

Platform	What it needs	Where to find it
Zendesk Guide	Subdomain. Optionally an API token and its email address.	The acme in acme.zendesk.com. Add the token if you want drafts to come across too.
Intercom	Access token, and a region if you're not in the US.	Developer Hub > your app > Configure > Authentication. Tokens from elsewhere in Intercom's settings are rejected. Set the region to eu or au if your workspace is hosted there. Drafts and translations come across as well as published articles.
Help Scout Docs	Docs API key.	Manage > Docs > API Keys in Help Scout.
Freshdesk	Subdomain and API key.	Your Freshdesk profile settings. Freshdesk's Categories and Folders both come across as HelpDocs folders.
Document360	API token.	Your Document360 settings.
GitBook	API token and space ID.	Group pages become folders and document pages become articles under them.
GitHub	Repository as owner/name, or the URL of your docs folder.	See the notes below this table.
Notion	Integration token, plus the pages shared with that integration.	Notion has no folders, so pages whose parent is another page are grouped under a folder named after the parent.
Helpjuice	Subdomain and API key.	Your Helpjuice account settings.
ProProfs	The URL of an export .zip.	Generate the export in ProProfs first.
ReadMe	A .zip export, uploaded directly.	Export from ReadMe, then drag the file in.
Google Drive	Your Google account, then pick the documents.	Each Google Doc becomes an article and immediate subfolders become folders. No AI, no credits.

A few notes on GitHub, which varies more than the rest:

• Set a docs folder. Without one, every Markdown file in the repository comes across, including the readme and the contributing guide.
• An access token is required for private repositories, and raises the rate limit on public ones.
• Branch, tag or commit defaults to the repository's default branch.
• Links between imported pages are rewritten to point at your new articles. Links to files outside the import are left as they are, and the review step tells you how many.
• Docusaurus and MkDocs ordering, like sidebar_position, isn't imported. HelpDocs orders content separately, so you set the order here.

From a file

CSV file. Any spreadsheet export. We read your headers, guess which HelpDocs field each column is, and let you correct the guesses. Full column reference in Import from a CSV.

Word & other docs. Up to 25 files: .docx, .rtf, .odt, .md, .html or .txt. We convert each one and split it into articles at its own headings, so your content comes through exactly as written. Choose to split at top-level headings, at sub-headings, or one article per file. A document with no headings comes in as a single article named after the file. Free, and instant.

With AI

Two sources use AI, because the content arrives with no structure to work from. Both spend AI credits.

Source	Cost	Limits
PDF files	10 credits per file	25 files, 15 MB each, 100 MB per import
Crawl a website	3 credits per page	50 pages per crawl

PDF files have no headings to split on, so AI reads each one and works out where the articles are.

Crawl a website is for a platform with no export and no API. Point us at a help site, we follow its links staying on the same site, and extract an article from each page. It's the roughest of the sources, since we're reading a rendered page rather than a structured export, so read the preview carefully. Each crawled article keeps its page URL, so redirects are set up and you can re-run the crawl later in Update existing mode to refresh what changed.

What happens to your content on the way in

• Headings are normalised. HelpDocs articles use heading levels 3 to 5, so a heading outside that range is moved into it. The article title is the top level, so an H1 in the body becomes an H3.
• Images are copied to HelpDocs. Anything the import wrote gets its images pulled across and rehosted, so your articles keep working after you switch the old site off. Only articles from this import are touched.
• Redirects are set up where we can. Zendesk, Intercom, Help Scout, Helpjuice, ProProfs, ReadMe, GitHub and the website crawl all tell us the old URL of each page, so we point it at the new article. A CSV can too, if you map its URL columns. Redirects from an earlier import are kept, so a second import doesn't undo the first.
• Links between imported articles are rewritten where the source lets us match them. Links pointing outside the import are left alone.
• Search is reindexed when the import finishes, so give it a minute before your new articles turn up in search.

Your original content is untouched

Importing copies. Your Zendesk, Notion or GitHub content stays exactly where it is, and you can run the import as many times as you like.

Undoing an import

For an hour after an import finishes, there's an Undo button next to it under Recent migrations at the bottom of the Migrations tab. It counts down, so you can see how long is left.

Undo removes everything that import created. If it was a Replace everything import, the content it deleted on the way in is put back too, so your knowledge base ends up exactly as it was before you started.

An hour is deliberately generous for the thing it's for. Most imports are either obviously right or obviously wrong the moment you look at them, and the hour is there for the ones where you want to click into a few articles and check the conversion before deciding.

Removing an import later

Once the hour is up the button becomes Delete imported content. It removes the articles and folders the import created, and it's an ordinary deletion rather than an undo: it doesn't put back anything a Replace everything import removed on its way in.

Because time has passed, it shows you what it's about to do first:

• How many articles and folders will go.
• How many of them you've edited since the import, listed first, because those edits go too.
• Any folders that are being kept, because they've gained other content since and deleting them would take that with it.

Nothing is deleted until you confirm against that list. Articles removed this way go to your deleted articles list, where you can restore them for a month before they're cleared for good.

If something goes wrong

• Stop a running import with the Cancel button. Anything already written stays put and is searchable, and nothing further is written.
• Back up first if you're nervous. The Export tab gives you a CSV or ZIP of everything you have before you start.
• Read the review notes. They name specific files and counts, like how many cross-links couldn't be resolved, and they're the fastest way to know what needs a look.
• Check a few articles afterwards, particularly ones with tables, code samples or embedded media, and do it while Undo is still available.

Limits

• One import can run at a time per account.
• Up to 10 imports can be started in any 10 minute window.
• Document and PDF uploads: 25 files per import, 15 MB per file, 100 MB in total.
• Website crawls: 50 pages per import.
• Undo is available for an hour after an import finishes.
• Importing is paused while an account is on an expired trial or past due.

Some of these importers are new

Freshdesk, Document360, GitBook, GitHub, Notion, Google Drive, the document and PDF uploads and the crawler are all recent, and Intercom was rebuilt to read the API rather than scrape your public help site. None of them can damage what you already have, since nothing is written until you review it and Undo removes anything you don't want. What's still settling is how faithfully every variation of a source converts, so read the preview before you import and check a few articles after.

Not sure where to begin?

If you're starting a knowledge base from scratch rather than moving one, we've got starter CSV templates for a few industries. Replace the content in [brackets] with your own and import the file.

SaaS 🚀

Common articles about account management, billing, and feature documentation. Download CSV here.

E-Commerce 🛍️

Templates for order tracking, returns, product information, and shipping policies. Download CSV here.

Education 🎓

Articles for course enrollment, technical support, and student resources. Download CSV here.

Logistics 🚚

Tracking information, shipping details, and common delivery questions. Download CSV here.

Still stuck?

If your platform isn't on the list, or an export isn't coming across the way you expect, get in touch through the widget and send us a sample. We'd rather build the importer than have you retype your knowledge base 😊