Confluence

The Confluence connector indexes pages — titles and full page bodies — into the local SQLite index and exposes read tools across pages, blog posts, comments, and spaces, plus a small set of HITL-gated write tools for creating and updating content. Confluence Cloud only.

What’s indexed

Service id: confluence
Item types: page
Body: the page’s storage-format content converted to plain text, up to 16,384 UTF-16 code units.

Indexing depth is enforced on every sync, not just at reindex time. See Indexing depth.

Authenticate

nimbus connector auth confluence

Stored in the Vault as confluence.base_url, confluence.email, and confluence.api_token and injected at spawn as CONFLUENCE_BASE_URL, CONFLUENCE_EMAIL, CONFLUENCE_API_TOKEN. The MCP server uses Atlassian Basic auth (email + API token). Generate an API token at id.atlassian.com → Security → API tokens.

Read tools (no HITL)

Tool	Purpose
`confluence_blogpost_get`	Get a blog post by id (`GET /content/{id}`).
`confluence_blogpost_list`	List blog posts in a space (`GET /content?type=blogpost`).
`confluence_comment_list`	List footer comments on a page (`GET /content/{id}/child/comment`).
`confluence_page_get`	Get a Confluence page with `body.storage` (`GET /content/{id}`).
`confluence_page_list`	List pages in a space (`GET /content?type=page`).
`confluence_space_list`	List Confluence spaces (`GET /space`).

Write tools (HITL-gated)

Every tool below requires interactive consent through the Gateway HITL gate before it executes.

Tool	Purpose
`confluence_comment_add`	Add a footer comment to a page.
`confluence_page_create`	Create a page in a space (optional `parentPageId`).
`confluence_page_update`	Update page body and bump version (pass current version + title).

Platform notes

Confluence Cloud only. The base URL must be the *.atlassian.net site root; the connector appends /wiki automatically.

See also