Confluence
Add Confluence Cloud pages, or a whole space, to the docs library. Limn reads them with an Atlassian API token, keeps them current, and never changes a page.
On this page
Many teams keep their specifications, runbooks and decisions in Confluence. Limn can add a Confluence page and every page below it, or a whole space, to the workspace’s docs library. From there, reviews, story checks and the docs chat can use them. Limn reads Confluence only; it never edits, comments on or creates a page.
#What you need
- Confluence Cloud (an
atlassian.netsite). - An Atlassian account email and API token, from id.atlassian.com → Security → API tokens. The token acts as you, so it reads only the spaces and pages you can see.
- AI set up in Settings → AI, because Docs needs it.
#Set it up
If Jira is this workspace’s tracker and its email and token are saved, a Confluence page on the same site uses the Jira connection, because Atlassian Cloud uses one API token for both. Settings → Docs then shows Using your Jira connection, and there is nothing to set up.
Otherwise, or to use a different token:
- Open Settings → Docs. Under Confluence, type the Site, your Email and the API token (or press Use a different token when the Jira connection is in use).
- Save, and press Test connection.
#Add pages
- On the Docs page, press + and choose Confluence. You can also paste a Confluence address into Web page or site; Limn notices and switches.
- Paste a page address. A space’s home page, or the space’s own address, adds the whole space; any other page adds it and every page below it, never its siblings or parents.
- Limn counts the pages first (“This will add 12 pages from Payments handbook.”). Press Start.
The pages become one document, with a chapter per page, up to 100 pages. For a bigger space, add its sections as separate documents. A short /wiki/x/… link does not name a page; open the page and copy its full address.
#What Limn reads
- The pages’ titles, their place in the page tree, their text and their version.
- Tables, code blocks, info and warning panels, expand sections, task lists, and status and Jira macros, as text.
- Images and diagrams: Mermaid and PlantUML as their source; draw.io and Gliffy as an outline where Limn can read one; others as a named placeholder.
- Navigation macros, such as a table of contents, are left out.
#How often Limn reads
Limn checks each Confluence document for changes about every three hours while the workspace’s library is in use, and just before a review or story check reads it. It asks Confluence which pages changed and reads only those, then summarises only the sections whose text changed. Press Refresh now in the document’s menu to check at once, or Rebuild to read every page again.
#What Limn writes
Nothing. Limn never edits, comments on or creates a Confluence page.
#Troubleshooting
| What you see | What to do |
|---|---|
| Add a Confluence API token in Settings | Save a token in Settings → Docs, or save Jira’s email and token if Jira is on the same site. |
| Can't reach this page in Confluence since … | The page moved, was deleted, or the token lost access. The last good version stays readable; fix access and press Refresh now. |
| Only the first 100 pages fit in one document. | Add the rest of the space as separate sections. |
Where the token is kept: Data and storage.