The custom website connector can currently only index HTML and text, see https://support.atlassian.com/organization-administration/docs/connect-custom-website-to-rovo/. Why isn't PDF and DOCX also supported here, as it is with other connectors, e.g. Dropbox or MS OneDrive? The technology is available.
Many guides and tutorials are available in HTML on our internal documentation system, but unfortunately not all of them. The HTML pages are indexed wonderfully and taken into account by Rovo with really good results. However, because the PDF documents are missing, the quality of Rovo varies greatly, which undermines employee confidence. For the average user, it is absolutely incomprehensible why the results for some products are correct, while others show completely outdated information from old Jira tickets.
It would be great if the custom website connector could also be expanded to include additional data types.
Kind regards,
Rainer
