The following is an example of simple configuration for Confluence generation and publishing:

extensions = ['sphinxcontrib.confluencebuilder']
confluence_publish = True
confluence_space_name = 'TEST'
confluence_parent_page = 'Documentation'
confluence_server_url = ''
confluence_server_user = 'username'
confluence_server_pass = 'password'

All Atlassian Confluence Builder configurations are prefixed with confluence_. View the entire list of configuration options below.

essential configuration


A boolean that decides whether or not to allow publishing. This option must be explicitly set to True if one wishes to publish content. By default, the value is set to False.

confluence_publish = True


The password value used to authenticate with the Confluence instance. If using Confluence Cloud, it is recommended to use an API token for the configured username value (see API tokens):

confluence_server_pass = 'vsUsrSZ6Z4kmrQMapSXBYkJh'

If API tokens are not being used, the plain password for the configured username value should be used:

confluence_server_pass = 'myawesomepassword'


It is never recommended to store an API token or raw password into a committed/shared repository holding documentation. A documentation’s configuration can modified various ways with Python to pull an authentication token for a publishing event (reading from a local file, acquiring a password from getpass, etc.). If desired, this extension provides a method for prompting for a password (see confluence_ask_password).


The URL for Confluence. The URL should be prefixed with https:// or http:// (depending on the URL target). The target API folder should not be included in the URL (i.e. excluding rest/api/). For a Confluence Cloud instance, an example URL configuration is as follows:

confluence_server_url = ''

For a Confluence Server instance, an example URL configuration, if the instance’s REST API is, should be as follows:

confluence_server_url = ''


The username value used to authenticate with the Confluence instance. If using Confluence Cloud, this value will most likely be the account’s E-mail address. If using Confluence instance, this value will most likely be the username value.

confluence_server_user = ''
confluence_server_user = 'myawesomeuser'


Key of the space in Confluence to be used to publish generated documents to.

confluence_space_name = 'MyAwesomeSpace'

Note that the space name can be case-sensitive in most (if not all) versions of Confluence.

generic configuration


New in version 1.2.

Add section numbers to page and section titles if doctree uses the :numbered: option. By default, this is enabled:

confluence_add_secnumbers = True

See also confluence_publish_prefix.


New in version 1.3.

Explicitly set which alignment type to use when a default alignment value is detected. As of Sphinx 2.0+, the default alignment is set to center. Legacy versions of Sphinx had a default alignment of left. By default, this extension will use a Sphinx-defined default alignment unless explicitly set by this configuration value. Accepted values are left, center or right.

confluence_default_alignment = 'left'


The name of the file to use header data. If provided, the contents found inside the header file will be added to the start of all generated documents. The file path provided should be relative to the build environment’s source directory. For example:

confluence_header_file = 'assets/header.tpl'

See also confluence_footer_file.


An integer value, if provided, to indicate the maximum depth permitted for a nested child page before its contents is inlined with a parent. The root of all pages is typically the configured master_doc. The root page is considered to be at a depth of zero. By defining a value of 0, all child pages of the root document will be merged into a single document. By default, the maximum document depth is disabled with a value of None.

confluence_max_doc_depth = 2


A boolean value to whether or not nest pages in a hierarchical ordered. The root of all pages is typically the configured master_doc. If a master_doc instance contains a toctree, listed documents will become child pages of the master_doc. This cycle continues for child pages with their own toctree markups. By default, the hierarchy mode is disabled with a value of False.

confluence_page_hierarchy = False

Note that even if hierarchy mode is enabled, the configured master_doc page and other published pages that are not defined in the complete toctree, these documents will still be published based off the configured (or unconfigured) confluence_parent_page setting.


New in version 1.2.

A string value to where to include previous/next buttons (if any) based on the detected order of documents to be included in processing. Values accepted are either bottom, both, top or None. By default, no previous/next links are generated with a value of None.

confluence_prev_next_buttons_location = 'top'


New in version 1.2.

The suffix to put after section numbers, before section name.

confluence_secnumber_suffix = '. '

See also confluence_add_secnumbers.

publishing configuration



User’s running Cygwin/MinGW may need to invoke with winpty to allow this feature to work.

Provides an override for an interactive shell to request publishing documents using an API key or password provided from the shell environment. While a password is typically defined in the option confluence_server_pass (either directly set/fetched from the project’s or passed via a command line argument -D confluence_server_pass=password), select environments may wish to provide a way to provide an authentication token without needing to modify documentation sources or having a visible password value in the interactive session requesting the publish event. By default, this option is disabled with a value of False.

confluence_ask_password = False

A user can request for a password prompt by invoking build event by passing the define through the command line:

sphinx-build [options] -D confluence_ask_password=1 <srcdir> <outdir>

Note that some shell sessions may not be able to pull the password value properly from the user. For example, Cygwin/MinGW may not be able to accept a password unless invoked with winpty.


New in version 1.2.

Provides an override for an interactive shell to request publishing documents using a user provided from the shell environment. While a user is typically defined in the option confluence_server_user, select environments may wish to provide a way to provide a user without needing to modify documentation sources. By default, this option is disabled with a value of False.

confluence_ask_user = False


A boolean value to explicitly disable the automatic generation of titles for documents which do not have a title set. When this extension processes a set of documents to publish, a document needs a title value to know which Confluence page to create/update. In the event where a title value cannot be extracted from a document, a title value will be automatically generated for the document. For automatically generated titles, the value will always be prefixed with autogen-. For users who wish to ignore pages which have no title, this option can be set to True. By default, this option is set to False.

confluence_disable_autogen_title = True

See also:


A boolean value which explicitly disable any page update notifications (i.e. treats page updates from a publish request as minor updates). By default, notifications are enabled with a value of False.

confluence_disable_notifications = True

See also confluence_watch.


New in version 1.3.

Defines a list of labels to apply to each document being published. When a publish event either adds a new page or updates an existing page, the labels defined in this option will be added/set on the page. For example:

confluence_global_labels = ['label-a', 'label-b']

For per-document labels, please consult the confluence_metadata directive. See also confluence_append_labels.


A boolean value to whether or not force the configured space’s homepage to be set to the page defined by the Sphinx configuration’s master_doc. By default, the master_doc configuration is ignored with a value of False.

confluence_master_homepage = False


The root page found inside the configured space (confluence_space_name) where published pages will be a descendant of. The parent page value is used to match with the title of an existing page. If this option is not provided, pages will be published to the root of the configured space. If the parent page cannot be found, the publish attempt will stop with an error message. For example, the following will publish documentation under the MyAwesomeDocs page:

confluence_parent_page = 'MyAwesomeDocs'

If a parent page is not set, consider using the confluence_master_homepage option as well. Note that the page’s name can be case-sensitive in most (if not all) versions of Confluence.


New in version 1.2.

If set, the postfix value is added to the title of all published documents. In Confluence, page names need to be unique for a space. A postfix can be set to either:

  • Add a unique naming schema to generated/published documents in a space which has manually created pages; or,
  • Allow multiple published sets of documentation, each with their own postfix value.

An example publish postfix is as follows:

confluence_publish_postfix = '-postfix'

By default, no postfix is used. See also confluence_publish_prefix.


If set, the prefix value is added to the title of all published documents. In Confluence, page names need to be unique for a space. A prefix can be set to either:

  • Add a unique naming schema to generated/published documents in a space which has manually created pages; or,
  • Allow multiple published sets of documentation, each with their own prefix value.

An example publish prefix is as follows:

confluence_publish_prefix = 'prefix-'

By default, no prefix is used. See also confluence_publish_postfix.



Publishing individual/subset of documents with this option may lead to unexpected results.

A boolean value to whether or not purge legacy pages detected in a space or parent page. By default, this value is set to False to indicate that no pages will be removed. If this configuration is set to True, detected pages in Confluence that do not match the set of published documents will be automatically removed. If the option confluence_parent_page is set, only pages which are a descendant of the configured parent page can be removed; otherwise, all pages in the configured space could be removed.

confluence_purge = False

While this capability is useful for updating a series of pages, it may lead to unexpected results when attempting to publish a single-page update. The purge operation will remove all pages that are not publish in the request. For example, if an original request publishes ten documents and purges excess documents, a following publish attempt with only one of the documents will purge the other nine pages.


A boolean value to which indicates that any purging attempt should be done from the root of a published master_doc page (instead of a configured parent page; i.e. confluence_parent_page). In specific publishing scenarios, a user may wish to publish multiple documentation sets based off a single parent/container page. To prevent any purging between multiple documentation sets, this option can be set to True. When generating legacy pages to be removed, this extension will only attempt to populate legacy pages based off the children of the master_doc page. This option still requires confluence_purge to be set to True before taking effect.

confluence_purge_from_master = False


New in version 1.3.

Allows a user to override the title value for a specific document. When documents are parsed for title values, the first title element’s content will be used as the publish page’s title. Select documents name not include a title and are ignored; or, documents may conflict with each other but there is a desire to keep them the same name in reStructuredText form. With confluence_title_overrides, a user can define a dictionary which will map a given docname into a title element instead of the title element (if any) found in the respective document. By default, documents will give assigned titles values based off the first detected title element with a value of None.

confluence_title_overrides = {
    'index': 'Index Override',

See also:


Force a timeout (in seconds) for network interaction. The timeout used by this extension is not explicitly configured (i.e. managed by Requests and other implementations). By default, assume that any network interaction will not timeout. Since the target Confluence instance is most likely to be found on an external server, is it recommended to explicitly configure a timeout value based on the environment being used. For example, to configure a timeout of ten seconds, the following can be used:

confluence_timeout = 10


New in version 1.3.

Indicate whether or not the user publishing content will automatically watch pages for changes. In Confluence, when creating a new page or updating an existing page, the editing user will automatically watch the page. Notifications on automatically published content is typically not relevant to publishers through this extension, especially if the content is volatile. If a publisher wishes to be keep informed on notification for published pages, this option can be set to True. By default, watching is disabled with a value of False.

confluence_watch = False

See also confluence_disable_notifications.

advanced publishing configuration


New in version 1.3.

Allows a user to decide how to manage labels for an update page. When a page update contains new labels to set, they can either be stacked on existing labels or replaced. In the event that a publisher wishes to replace any existing labels that are set on published pages, this option can be set to False. By default, labels are always appended with a value of True.

confluence_append_labels = True

See also confluence_global_labels.


Provides an override for asset publishing to allow a user publishing to either force re-publishing assets or disable asset publishing. This extension will attempt to publish assets (images, downloads, etc.) to pages via Confluence’s attachment feature. Attachments are assigned a comment value with a hash value of a published asset. If another publishing event occurs, the hash value is checked before attempting to re-publish an asset. In unique scenarios, are use may wish to override this ability. By configuring this option to True, this extension will always publish asset files (whether or not an attachment with a matching hash exists). By configuring this option to False, no assets will be published by this extension. By default, this automatic asset publishing occurs with a value of None.

confluence_asset_override = None


Provide a CA certificate to use for server certificate authentication. The value for this option can either be a file of a certificate or a path pointing to an OpenSSL-prepared directory. If configured to use REST API (default), refer to the Requests SSL Cert Verification documentation (verify) for information. If server verification is explicitly disabled (see confluence_disable_ssl_validation), this option is ignored. By default, this option is ignored with a value of None.

confluence_ca_cert = 'ca.crt'


Provide a client certificate to use for two-way TLS/SSL authentication. The value for this option can either be a file (containing a certificate and private key) or as a tuple where both certificate and private keys are explicitly provided. If a private key is protected with a passphrase, a user publishing a documentation set will be prompted for a password (see also confluence_client_cert_pass). By default, this option is ignored with a value of None.

confluence_client_cert = 'cert_and_key.pem'
# or
confluence_client_cert = ('client.cert', 'client.key')


Provide a passphrase for confluence_client_cert. This prevents a user from being prompted to enter a passphrase for a private key when publishing. If a configured private key is not protected by a passphrase, this value will be ignored. By default, this option is ignored with a value of None.

confluence_client_cert_pass = 'passphrase'



It is not recommended to use this option.

A boolean value to explicitly disable verification of server SSL certificates when making a publish request. By default, this option is set to False.

confluence_disable_ssl_validation = False


The page identifier check for confluence_parent_page. By providing an identifier of the parent page, both the parent page’s name and identifier must match before this extension will publish any content to a Confluence instance. This serves as a sanity-check configuration for the cautious.

confluence_parent_page_id_check = 1


REST calls use the Requests library which will use system-defined proxy configuration; however, a user can override the system-defined proxy by providing a proxy server using this configuration.

confluence_proxy = 'myawesomeproxy:8080'


New in version 1.3.

When a user wishes to start managing a new document set for publishing, there maybe concerns about conflicts with existing content. When the dry run feature is enabled to True, a publish event will not edit or remove any existing content. Instead, the extension will inform the user which pages will be created, whether or not pages will be moved and whether or not pages/attachments will be removed. By default, the dry run feature is disabled with a value of False.

confluence_publish_dryrun = True

See also Confluence Spaces and Unique Page Names.


New in version 1.3.

A publish event will from this extension will typically upload new pages or update existing pages on future attempts. In select cases, a user may not wish to modify existing pages and only permit adding new content to a Confluence space. To achieve this, a user can enable a “only-new” flag which prevents the modification of existing content. This includes the restriction of updating existing pages/attachments as well as deleting content as well. By default, the only-new feature is disabled with a value of False.

confluence_publish_onlynew = True



If confluence_publish_subset is configured, this option disables confluence_purge.

Provides the ability for a publisher to explicitly list a subset of documents to be published to a Confluence instance. When a user invokes sphinx-build, a user has the ability to process all documents (by default) or specifying individual filenames which use the provide files and detected dependencies. If the Sphinx-detected set of documents to process contain undesired documents to publish, confluence_publish_subset can be used to override this. Defined document names should be a relative file path without the file extension. For example:

confluence_publish_subset = ['index', 'foo/bar']

A user can force a publishing subset through the command line:

sphinx-build [options] -D confluence_publish_subset=index,foo/bar \
    <srcdir> <outdir> index.rst foo/bar.rst

By default, this option is ignored with a value of []. See also manage publishing a document subset.


An authentication handler which can be directly provided to a REST API request. REST calls in this extension use the Requests library, which provide various methods for a client to perform authentication. While this extension already provided simple authentication support (via confluence_server_user and confluence_server_pass), a publisher may need to configure an advanced authentication handler to support a target Confluence instance.

Note that this extension does not define custom authentication handlers. This configuration is a passthrough option only. For more details on various ways to use authentication handlers, please consult Requests – Authentication. By default, no custom authentication handler is provided to generated REST API requests (if any).

from requests_oauthlib import OAuth1


confluence_server_auth = OAuth1(client_key,


A dictionary value which allows a user to pass key-value cookie information for authentication purposes. This is useful for users who need to authenticate with a single sign-on (SSO) provider to access a target Confluence instance. By default, no cookies are set with a value of None.

confluence_server_cookies = {
    'SESSION_ID': '<session id string>',
    'U_ID': '<username>'

advanced processing configuration


The file name suffix to use for all generated files. By default, all generated files will use the extension .conf (see confluence_file_transform).

confluence_file_suffix = '.conf'


A function to override the translation of a document name to a filename. The provided function is used to perform translations for both Sphinx’s get_outdated_docs and write_doc methods. The default translation will be the combination of “docname + confluence_file_suffix”.


New in version 1.2.

Provides a dictionary of named JIRA servers to reference when using the jira or jira_issue directives. In a typical Confluence environment which is linked with a JIRA instance, users do not need to take advantage of this configuration – Confluence should automatically be able to link to respectively JIRA issues or map JIRA query languages with a configured JIRA instance. In select cases where an instance has more than one JIRA instance attached, a user may need to explicitly reference a JIRA instance to properly render a JIRA macro. JIRA-related directives have the ability to reference JIRA instances, with a combination of a UUID and name; for example:

.. jira_issue:: TEST-151
    :server-id: d005bcc2-ca4e-4065-8ce8-49ff5ac5857d
    :server-name: MyAwesomeJiraServer

It may be tedious for some projects to add this information in each document. As an alternative, a configuration can define JIRA instance information inside a configuration option as follows:

confluence_jira_servers = {
    'server-1': {
        'id': '<UUID of JIRA Instance>',
        'name': '<Name of JIRA Instance>'

With the above option defined in a project’s configuration, the following can be used instance inside a document:

.. jira_issue:: TEST-151
    :server: server-1


A function to override the translation of literal block-based directive language values to Confluence-support code block macro language values. The default translation accepts Pygments documented language types to Confluence-supported syntax highlight languages.

def my_language_translation(lang):
    return 'default'

confluence_lang_transform = my_language_translation


A boolean value to whether or not automatically remove the title section from all published pages. In Confluence, page names are already presented at the top. With this option enabled, this reduces having two leading headers with the document’s title. In some cases, a user may wish to not remove titles when custom prefixes or other custom modifications are in play. By default, this option is enabled with a value of True.

confluence_remove_title = True

See also: