Pages created via API are changed ("p" tags added)

[Hleb Valoshka]

Good day.

We're migrating on-premise Confluence to cloud. Some of our pages are created via REST API during application rollouts. With on-premise instance we were using v1 API, with cloud one we use v2.

Pages pushed to cloud are broken, they have additional "p" tags inserted. For example, the old page:

<ul>
          <li>Redacted<ul>
            <li>
             <a class="external-link" href="redacted" rel="nofollow">redacted</a><i> (redacted)</i>
            </li><li>
             <a class="external-link" href="redacted" rel="nofollow">redacted</a><i> (redacted)</i>
             
            </li>
           </ul>
 ...

While the new one is:

<ul class="ak-ul" data-indent-level="1"><li><p data-renderer-start-pos="720">Redacted</p><ul class="ak-ul" data-indent-level="2"><li><p data-renderer-start-pos="752"><a data-testid="link-with-safety" href="redacted" title="redacted" data-renderer-mark="true" class="cc-1rn59kg">redacted</a><em data-renderer-mark="true"> (redacted)</em>
            </p></li><li><p data-renderer-start-pos="799"><a data-testid="link-with-safety" href="redacted" title="redacted" data-renderer-mark="true" class="cc-1rn59kg">redacted</a><em data-renderer-mark="true"> (redacted)</em>
             
            </p></li></ul>

How to make Confluence not to insert any unrequested content?

[marc -Collabello--Phase Locked-]

Not sure, but it looks like you are not looking at the Confluence storage format, but at the rendered html format.

[Hleb Valoshka]

Yes, in on-premise Confluence we use rendered xml because we use macro. I don't think that undocumented json storage format is a proper replacement for that.

I also don't understand why Confluence can't simply show what it has but instead it transforms xml into json and then into html again.