Confluence KB Integration

Credentials required

Step 1 — Create an API token

1. 1
   Open API token management

   Go to https://id.atlassian.com/manage-profile/security/api-tokens and click Create API token.
2. 2
   Name and copy the token

   Give it a label (e.g. TR-AI KB indexer) and copy the token value — it is shown only once.

Step 2 — Find your Space Key

In Confluence, go to your space → Space Settings → Space Details. The Space Key is the short identifier (e.g. ENG, KB, SUPPORT). It appears in the URL when browsing the space: https://yoursite.atlassian.net/wiki/spaces/{SPACEKEY}/pages.

Step 3 — Enter credentials in TR-AI

In your TR-AI Dashboard, open the Knowledge Base section and select Confluence. Enter your Base URL, Email, API Token, and Space Key. Click Verify & Index. TR-AI will:

• Call GET /wiki/api/v2/spaces?keys={spaceKey}&limit=1 to verify credentials and confirm the space exists.
• Cursor-paginate all pages in the space (250 per page).
• Fetch each page body in parallel (concurrency 6) using GET /wiki/api/v2/pages/{id}?body-format=storage.
• Convert Confluence storage XML to plain text and chunk it for embedding.

How TR-AI fetches pages

Page discovery

TR-AI uses cursor-based pagination on GET /wiki/api/v2/spaces/{spaceKey}/pages with limit=250&status=current. It follows the _links.next cursor until all pages are discovered — there is no hard cap.

Content extraction

Page bodies are fetched in storage format (Confluence's XHTML-based storage format). TR-AI strips the XML to plain text, handling:

• Structured macros (ac:structured-macro) — inner text preserved.
• Block elements (p, h1–h6, li, tr, td) → newlines.
• HTML entities (&, <, >,  ) → decoded.
• Pages with fewer than 20 characters after stripping are skipped.

Credential verification errors

Re-indexing

To re-index after updating your Confluence space, click Re-index in the TR-AI Dashboard. You can also trigger a re-index via the API:

POST /api/admin/reindex
Content-Type: application/json

{ "trialAccountId": 123 }