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

# Confluence Cloud

> Confluence Cloud space ingestion into Flow knowledge search.

## What Flow indexes from Confluence Cloud

Confluence Cloud syncs pages from selected sites and spaces into Flow knowledge search.

When you connect, you name the Confluence site and approve Confluence read access. When you add a space or page tree as a source, you choose **Who can read this in Flow?** for that source. Each source on the connection makes its own choice.

## Prerequisites

* A Confluence Cloud account that can read the spaces you want to sync. Your organization may require an administrator to approve the connection.
* A Flow admin or knowledge manager who can manage knowledge sources under **Integrations**.
* The Confluence sites and spaces the team needs.

An Atlassian organization admin who can create a scoped organization API key is needed only for sources that copy Confluence permissions. You can attach that key when you connect or later from the connection. Sources with a Flow audience do not use it.

## Who can read this in Flow

When you add a source, Flow asks who can read the indexed pages. The four choices match the labels in the source settings. When a choice lists Flow teams, you select the teams; the team adding the source starts selected, and only an organization admin can select other teams.

| Choice | What Flow does |
| - | - |
| **People who can already open it in Confluence** (recommended) | Flow copies Confluence space and page permissions on every crawl, including restricted pages. Reads follow the Confluence permissions Flow crawled. Adding the source does not grant read. Until an Atlassian Admin connection is attached, the source shows **Waiting for Atlassian Admin** and nobody can read the indexed copy. |
| **People who can open it in Confluence, plus selected teams** | Flow copies Confluence permissions and also lets current members of the selected teams read the indexed copy, even when Confluence would deny them. You must confirm that before you continue. The Confluence part waits for an Atlassian Admin connection; the teams can read right away. |
| **Only selected teams** | Current members of the selected teams can read the indexed copy, even when they cannot open the page in Confluence. Restricted pages are skipped. No Atlassian Admin connection is needed. You must confirm that before you continue. |
| **Everyone in the organization** | Every current member of the Flow organization can read the indexed copy, including people outside the selected teams. Restricted pages are skipped. No Atlassian Admin connection is needed. Only an organization admin can choose this, and must confirm organization-wide access before continuing. |

By default, a source with a Flow audience skips a page when Confluence restricts who can read it. It also skips the page when an ancestor is restricted, or when Flow cannot prove the space is open to all licensed users. The optional source settings below explain when you can choose to share more content. Flow also skips a page when it could not read all required evidence, for example when Confluence returned a restriction subject Flow cannot identify. Those pages do not appear in **Knowledge**.

Use a dedicated low-privilege Atlassian account when a source has a Flow audience. The crawl can read everything that account can read.

### Include pages from spaces with limited access

This setting is off by default and is available when you share a source with a Flow team or organization. Some Confluence spaces are shared with only certain people or groups. Turn it on only if everyone selected in **Who can read this in Flow?** may read pages from those spaces, even if they cannot open them in Confluence. Flow asks you to confirm this before applying the change.

Flow still skips pages with their own viewing restrictions, restrictions from a parent page, or permission checks it could not complete. This setting does not change permissions in Confluence. Keep **Include restricted pages** off if you want those pages to stay out of Flow.

### Include restricted pages

A source with a Flow audience has an **Include restricted pages** setting, off by default. Turning it on stops the skip rule for that source and re-crawls it. Every page the connecting Atlassian account can read in that source then becomes visible to the chosen Flow audience, including pages that are restricted in Confluence. Flow asks you to confirm that before it applies the change. Turning it off again removes the restricted pages on the next crawl.

When more than one team shares a Confluence space, one team's source runs the crawl for all of them. That team can turn on **Include restricted pages** only after every other team sharing the space has turned it on for its own source. Each team decides for itself.

### Change who can read a source

Open the source and change **Who can read this in Flow?** in its settings. The change applies to that source only and takes effect on its next crawl. Choosing **Everyone in the organization**, or any choice that lets a team read pages Confluence would deny, requires the same confirmation as setup.

Flow starts a crawl right away only when the change affects the Confluence permissions Flow copies. That happens when a source starts copying Confluence permissions on a connection that already has an Atlassian Admin connection. It also happens when you attach an Atlassian Admin connection to a connection with sources that were waiting for it. Switching between **Only selected teams** and **Everyone in the organization** applies at once without a crawl.

## Prepare the provider

Confirm that the connected account can read the selected content. If your organization requires approval for connections, ask your Confluence administrator to approve it.

If any source will copy Confluence permissions, create the organization API key. You can attach it when you connect or later from the connection's **Knowledge** tab. Sources with a Flow audience never use it.

## Create the organization API key

Restricted Confluence spaces name people and groups with Atlassian account IDs. Flow uses an Atlassian organization API key to map those IDs to people in Flow.

Create this key only when a source copies Confluence permissions. Only an Atlassian organization admin can create it. Create a key with scopes. Do not create an unscoped key. An unscoped key can call every Atlassian Admin API.

<Steps>
  <Step title="Open Atlassian Administration.">
    Sign in at [Atlassian Administration](https://admin.atlassian.com/). If you have more than one organization, select the one that owns the Confluence site.
  </Step>

  <Step title="Create a scoped API key.">
    Open **Organization settings**, then **API keys**. Choose **API keys with scopes**, then **Create API key**. Give the key a name you will recognize later, such as Flow knowledge.
  </Step>

  <Step title="Set an expiry and the required scopes.">
    Atlassian caps organization API keys at one year. Set an expiry you will remember, and rotate the key before it ends. Select these scopes:

    * `read:workspaces:admin`
    * `read:directories:admin`
    * `read:accounts:admin`

    Then create the key and copy the secret. Atlassian shows the secret once.
  </Step>

  <Step title="Copy the organization ID.">
    The organization ID is the segment between `/o/` and the next `/` in the Administration address. Example:

    `https://admin.atlassian.com/o/`**`a1b2c3d4-0000-4000-8000-000000000000`**`/overview`

    Paste only `a1b2c3d4-0000-4000-8000-000000000000`. Do not paste the full address or the Confluence site URL. Atlassian may also show the organization ID on the create-key screen. If you see it there, copy both values in one visit. Flow does not derive the organization ID.
  </Step>
</Steps>

Atlassian documents organization API keys and scoped keys in [Manage an organization with the admin APIs](https://support.atlassian.com/organization-administration/docs/manage-an-organization-with-the-admin-apis/). The [Organizations REST API introduction](https://developer.atlassian.com/cloud/admin/organization/rest/intro/) also tells administrators to copy the organization ID with the key.

## Connect in Flow

Open **Integrations**, choose **Confluence Cloud**, and complete the connection flow.

<Steps>
  <Step title="Enter the Confluence site name.">
    Flow shows it as `https://<site>.atlassian.net`. Use only letters, digits, and hyphens, and do not start or end with a hyphen.

    Optionally choose an **Atlassian Admin connection**. If you have not connected Atlassian Admin yet, choose **Attach later**. See [Create the organization API key](#create-the-organization-api-key).
  </Step>

  <Step title="Approve Confluence read access for that site.">
    Use a dedicated low-privilege Atlassian account when sources will have a Flow audience.
  </Step>
</Steps>

Then add a source: select the team and the spaces or page trees to sync, and choose **Who can read this in Flow?**.

To attach the Atlassian Admin connection later, open the Confluence connection, then its **Knowledge** tab, and choose the Atlassian Admin connection. Sources that were waiting for it crawl again and become readable.

## Capabilities

* Select sites and spaces for one team.
* Choose Confluence permissions or a Flow audience for each source.
* Run scheduled or on-demand sync.
* Search the synced pages in Flow.

A team connection is how you select and sync the space. When a source copies Confluence permissions, read access follows the permissions Flow crawled, not team membership. Adding the source does not grant read, and it is not required for org-wide Knowledge or search.

When a source has a Flow audience, that audience is the read grant. Restricted pages are skipped.

## Automation integration reference

Automation action, scope, and risk reference lives in the [Confluence Cloud integration](/automations/connectors/confluence-cloud).

## Sync and behavior

If Flow restarts while checking access, it retries that work. A restart does not mean that your Confluence permissions are wrong. If a page remains excluded because Flow could not check its access, sync again or contact Flow support.

After Flow reads the selected spaces, it prepares eligible content for search. Wait for the sync and indexing to finish before checking that your team can find the content. If progress stops, contact Flow support; changing Confluence permissions does not fix a processing delay.

After the first sync, confirm that the intended pages appear under **Knowledge**.

By default, a sync for a source with a Flow audience includes only unrestricted pages in spaces open to all licensed users. **Include pages from spaces with limited access** also permits spaces shared with specific people or groups. Pages with their own or inherited viewing restrictions still stay out unless you separately enable **Include restricted pages**.

If a sync for a source with a Flow audience stops after repeated provider errors, Flow stops serving that source until a later sync completes. The pages stay indexed; they become readable again when the next sync confirms that they are still open.

In skipped-file lists, **Excluded by sharing settings** means the page does not meet the selected sharing policy. **Access check incomplete** means Flow could not finish checking whether the page is eligible; sync again or contact Flow support if it continues.

Flow can explain exclusions only for content the connection can discover. If Confluence hides a space or page entirely from that account, Flow cannot list it or count it as locked content. A successful sync therefore covers the selected content available to the connection, not every page in the site.

Later syncs recheck the selected pages and their access. They can still take time even when you have not edited any pages. Flow can reuse a saved copy of unchanged content to reduce requests to Confluence. It still checks whether the content is eligible for your team. If Confluence returns only page details without the text, Flow uses a verified saved copy of that version or requests the page text before processing it.

Flow may reuse saved permission and page hierarchy information for up to 30 minutes by default. Permission changes may take time to appear because Flow must check and apply them again.

## Attachment formats

Flow extracts attachments in PDF, Word (`.docx`), and PowerPoint (`.pptx`) formats. Other file types, including images and spreadsheets, appear as unsupported and are skipped before download. To make their content searchable, save it in a supported format that contains text. Changing Confluence permissions does not add support for a file type.

### An attachment cannot be read

A file can be listed as a supported format even when its contents use a different format or are damaged. Flow leaves that attachment out of search and shows the reading problem in the crawl report.

Open the file in its original application and export a new PDF, Word (`.docx`), or PowerPoint (`.pptx`) copy with selectable text. Replace the attachment in Confluence and sync again. If a valid supported file still fails, contact Flow support. Changing the connected account's permissions does not repair a file-format problem.

## Troubleshooting

| Symptom | What to check |
| - | - |
| Connection fails | Confirm the connected account can read the selected spaces and that your organization permits the connection. The Atlassian organization API key is not needed to connect. |
| Documents excluded by sharing settings | Flow found access restrictions on these pages or their spaces. Your selected sharing settings keep them out of search. Review who should find them before changing those settings. |
| Documents awaiting access verification | Flow could not finish checking access, so these pages stay out of search. Sync again. If this continues, contact Flow support before changing who can read the pages. |
| Confluence returned no page text | Open the page in Confluence. If it is empty, there is no text for Flow to index. If it contains text, contact Flow support so we can investigate why Confluence did not return it. This message does not mean that you need broader permissions. |
| File is larger than Flow can index | Split the file into smaller files or exclude it from the source. Flow stops processing files that exceed its size or page-count limits. Changing Confluence permissions will not resolve this limit. |
| A page is archived | Flow excludes pages that Confluence marks as archived. Ask the content owner to restore the page if it is still needed, then sync again. Archived pages do not require administrator credentials or repeated retries. |
| An archived attachment is missing | Flow skips attachments that Confluence identifies as archived. If you need the file in Flow, ask the content owner whether to restore it, then sync again. |
| An attachment is unavailable | Confluence did not make the file available to the connection. It may have been removed or its access may have changed. This does not confirm deletion. Ask the content owner to check it, then sync again after it is available. |
| Page rechecks are waiting for request capacity | Flow waits until it can make more requests, then retries automatically. This wait does not use up the normal retry attempts. You do not need broader Confluence permissions. If it remains stuck, contact Flow support. |
| A page is pending after an earlier error | Flow still has work queued to process or recheck the page. Wait for that work to finish before changing the connection. The earlier error may remain visible while the recheck runs. |
| Sync is waiting after a temporary internal error | Flow retries automatically while attempts remain. You do not need to change Confluence permissions. Check the source sync status for progress. |
| Flow cannot reach the content service | Flow retries temporary connection problems automatically while attempts remain. You do not need to change Confluence permissions. If sync stops or the problem continues, contact Flow support. |
| Sync pauses after a service restart | Flow normally resumes automatically. It may wait a few minutes before continuing safely. Check for progress; if it remains stuck, contact Flow support. You do not need to reconnect Confluence. |
| Sync has stopped after an internal error | Contact Flow support. Waiting does not restart a stopped sync. After the problem is resolved, run one manual sync and wait for it to complete. |
| Organization ID is rejected | Paste only the ID segment from `admin.atlassian.com/o/<organization-id>/...`, not the full address or the site URL. Flow does not derive this ID. |
| A source shows **Waiting for Atlassian Admin** | The source copies Confluence permissions, but the connection has no Atlassian Admin connection yet, so nobody can read the indexed copy. Attach one from the connection's **Knowledge** tab, or give the source a Flow audience in its settings. |
| A restricted page is missing from a source with a Flow audience | Expected. Flow skips restricted pages when the source has a Flow audience. Turn on **Include restricted pages** for that source if the whole audience may read them. |
| A source is not readable after changing who can read it | If the change starts copying Confluence permissions, the source is readable after its next crawl finishes. Check the source's sync status; a failed sync keeps it unreadable until a later sync completes. |
| A source with a Flow audience stopped returning results | Its last sync may have failed after repeated provider errors. Run a manual sync; the pages become readable again when it completes. |
| Automation action fails | Confirm the connected account still has access to the target provider record, group, workspace, or project. |
| Permission error | Reconnect the provider after updating provider-side roles or scopes. |

## Related topics

* [Knowledge overview](/integrations/knowledge/overview)
* [Search](/integrations/knowledge/search)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.