Swaru

Help

Adding your docs

The chat answers from your knowledge base and your documentation site: point it at a docs site and it keeps up as your docs change.

Articles you write or import are in your knowledge base. This guide is about reading a docs site you already have.

How it reads your docs

Swaru starts at the address you give and follows links to pages under it. From each page it keeps the title, the description and the main text. Navigation, headers, footers and scripts are left out. On a site of 10 pages or more, lines that appear on at least half the pages are dropped, since they say nothing about any one page.

Each page becomes an article the chat can search and link to. Your docs are read again about once a day. If a crawl finds nothing (say, your site is down), the last good copy is kept.

Set it up

  1. On the Chat page, under Words, look and rules, an admin fills in Docs site: where your docs start, such as https://docs.example.com/.
  2. To keep the crawl inside one part of your site, list the https address prefixes under Only pages under, one per line. Empty means the docs site address.
  3. Max pages is 200 unless you change it, and at most 500.
  4. Answer links into sets where the chat's links may point. Empty means your docs site.
  5. Press Save, then Crawl the docs now under Your docs.
  6. Refresh after a minute. Your docs shows how many pages were read and when.

Anyone on your team can press Crawl the docs now. Only admins change the settings.

What it can read

  • Public pages over https. Pages that need a sign-in can't be read.
  • Pages that can be reached by following links from the start page. A page nothing links to is not found.
  • HTML pages. PDFs, images and other files are skipped.
  • Text that is in the page as your server sends it. A page that only shows its text after JavaScript runs looks empty to the crawl.
  • Pages up to 2 MB. Pages with almost no text are skipped.
  • Addresses without their query string (the part after ?), so pages that differ only by it are read once.
  • Redirects, only when they stay inside your docs addresses.

The crawler identifies itself as Swaru-DocsCrawler. If your site blocks unknown bots, allow it.

After you change your docs

Press Crawl the docs now when you publish a change, rather than waiting for the daily crawl. To see what the change fixed, use Ask them again now on the Insights page: see Gaps in your docs.

Still stuck? Write to hello@getswaru.com.