> ## Documentation Index
> Fetch the complete documentation index at: https://docs.truecourse.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Confluence

> Add any space of a connected Atlassian site as a source. Every current page becomes a document.

<Note>
  [Connections](/enterprise/connections) is an [enterprise](/enterprise/overview) feature. An open build has no Connections tab.
</Note>

A Confluence **source** is one space, added on [Context](/guard/web-sources) like any other source. It reads through the workspace's Atlassian connection, and one connection can serve several spaces.

## Connect the site

Confluence reads through the workspace's **Atlassian** connection: the site URL, an account email and an API token, made once in **Settings › Connections** and shared with Jira. See [Connections](/enterprise/connections#connect-an-atlassian-site) for where each value comes from. The account needs read access to the spaces you will add.

## Add a space as a source

Press **Add context** on any Context page. Once the site is connected, **Confluence** is offered among the connected accounts.

<Steps>
  <Step title="Scope">
    Enter the **Space key**, the short identifier of the space, usually a few uppercase letters. Two ways to find it:

    * From a page's URL, it is the segment after `/spaces/`: in `https://…/wiki/spaces/ENG/pages/…` the key is `ENG`.
    * From **Space settings › Space details**, the **Key** field.

    The source reads every current page of the space. Pages that carry no specification content (meeting notes, templates, status reports) are set aside by the [Document scan](/guard/spec-scan), so you do not need to clean the space up first.
  </Step>

  <Step title="Check, then add">
    **Check** counts the pages of the space and lists the first few, storing nothing. A wrong key surfaces here as *No space with key …*. Then press **Add and sync**.
  </Step>
</Steps>

The source is titled after the space and syncs like a documentation site: when it is added, on **Sync now**, and once a day.

## What a page becomes

One page is one document, titled after the page, with its body converted to markdown (headings kept, so the scan can slice it) and a deep link back to Confluence. Ahead of the body the document states when the page was created, when it was last changed, and its version number. The scan reads the dates, so when two documents disagree it knows which one is newer. A page whose text has not changed is never re-curated.

## Not read

Attachments, comments, and pages in the trash or in draft are not part of the document. Confluence Server and Data Center are not supported; the connection is Confluence Cloud with an API token.