GroundTruth
Back to Docs
Integration

Confluence

Import your Confluence Cloud pages into your GroundTruth knowledge base so every verification check runs against your live internal documentation.

This is a read-only integration. GroundTruth only reads your Confluence pages to build your knowledge base — it never modifies, deletes, or writes anything back to Confluence.

How it works

  1. Connect — Go to Settings and click "Connect" next to Confluence. You'll be redirected to authorize via OAuth.
  2. Sync — Once connected, click "Sync" to pull pages into your knowledge base. Documents appear in your Library with source "confluence".
  3. Verify — All future POST /api/check calls automatically search synced pages for evidence.

Prerequisites

  • A Confluence Cloud instance with an Atlassian OAuth 2.0 app.
  • The app needs the read:page:confluence and read:content-details:confluence granular scopes.

What syncs

All pages from your Confluence space (up to 500). Page bodies are extracted from Confluence storage format and converted to plain text. Content is stored (50 KB preview) for search.

Deduplication: Re-syncing updates existing documents via content hashing instead of creating duplicates.

Mapping: Each Confluence page becomes one document in your Library. The page title becomes the document title.

Token management

OAuth tokens are refreshed automatically before each sync. If a refresh fails, the existing token is used as a fallback. If your connection stops working, disconnect and reconnect to obtain a fresh refresh token.

API alternative

Connect and sync programmatically instead of using the dashboard.

api
# Connect via API
POST /api/integrations
{ "provider": "confluence" }

# Trigger a sync
POST /api/integrations/sync
{ "provider": "confluence" }

# Response
{
  "success": true,
  "provider": "confluence",
  "documentsFound": 142,
  "documentsIngested": 142
}

Data privacy

GroundTruth requests read-only permissions from Confluence. We only use the read:page:confluence and read:content-details:confluence scopes.

Your Confluence data is used solely to build your knowledge base for verification. It is never shared, modified, or used for model training.