Documentation

Everything you need to install Evergreen AI, run your first audit, and act on what it finds — for admins and page owners alike.

Evergreen AI for Confluence semantically audits your knowledge base with Atlassian-hosted AI and surfaces the pages most likely to be wrong — outdated facts, contradictions, deprecated references — each with quoted evidence and a confidence score. These pages explain how to use it end to end.

New here? The 30-minute path. Install from the Atlassian Marketplace, pick one space, and run the 100-page sample scan. You’ll have your first evidence-backed findings in minutes. Start with Getting started.

Browse the docs

Getting started

Install the app, review the permissions you approve, and run your first sample scan with the onboarding wizard.

Understanding findings

What a finding is, the six finding types, how confidence and evidence work, and why an empty report is a feature.

Health dashboard

The admin surface: space health scores, the findings list, filters and search, triage actions, and the scan ledger.

Page panel

What page owners and readers see on the page itself — audit status, findings with evidence, and one-click actions.

Scans & scheduling

Choose what gets scanned, schedule weekly or monthly runs, and understand incremental scanning and coverage.

Settings & budget

Scope, model tiers, sensitivity thresholds, the suppression list, the monthly budget, and data controls.

FAQ & troubleshooting

Common questions about data handling, pricing, and AI — plus what to do when a scan pauses or a page can’t be analyzed.

Key ideas in one minute

Looking for legal and data-handling detail? See Privacy, Security & data handling, the DPA, and Terms.