DocumentationIntegrationsopenDesk
openDesk knowledge
anymize full-text searches the XWiki of your openDesk instance and reads individual pages.
3 Min
What you can do with it
- “Search the wiki for the records-management instruction.”
- “What do we have on the topic of leave requests?”
- “Read the page Verwaltung.Onboarding and summarize it.”
- “Is there a page on procurement below the thresholds?”
Unlike with files, search here is full text, so over title and page content. The assistant searches first and then reads the page that fits best. Very long pages are shortened.
Which credentials do you need for XWiki?
There is no API key for openDesk knowledge. anymize signs in with username and password at the XWiki address of your instance, usually wiki.your-domain.
- 01Use the same username and password you use to sign in to the wiki.
- 02First check in the browser that sign-in with a password works, not only via SSO.
- 03If your wiki is not called xwiki, enter the wiki name in the optional field.
If your XWiki only allows SSO, this module stays off. The other openDesk modules are independent of that.
Connect
- 01Open the openDesk dialog and enter your domain.
- 02Turn on the Knowledge section.
- 03Enter username and password for XWiki. If your wiki is not called xwiki, you can give the name in the optional field.
- 04If the address differs from wiki.your-domain, enter it under “Address differs?”.
If your XWiki only allows sign-in via SSO, this module stays off. The other modules connect independently of that.
Do I still need a knowledge base in anymize then?
Those are two different things. The openDesk wiki stays where it is and is read when needed. A knowledge base in anymize takes documents in, prepares them and is designed for looking things up across many pages.
- Wiki: always current, nothing is copied, good for targeted follow-ups.
- Knowledge base: good when many documents should be searched together, including PDF and Word.
Read, do not write
There is no tool that creates or changes a wiki page. The connector only reads.
A service instruction often also contains names, extensions and file references. They are anonymized before the page goes to an AI model, and put back in the answer.