Privacy Policy
Evergreen AI for Confluence (“Evergreen AI,” “the app,” “we”) is an Atlassian Forge app built and operated by Andrew Murray le Roux, an independent developer (“Murray le Roux”). This policy explains what data the app processes, where it is stored, how long it is kept, and how it is deleted. It covers the app itself; the underlying Atlassian Cloud and Forge platform are governed by Atlassian’s own terms and privacy notices.
The short version: no data leaves Atlassian. The app declares no egress permissions. All processing — including AI inference — happens inside Atlassian’s environment, and all app data is stored in Atlassian-hosted Forge storage. There are no external servers, no third-party analytics, and no API keys.
Who is responsible
The app is provided by Andrew Murray le Roux (“Murray le Roux”), an independent single-developer vendor. For any privacy question or request, contact support@evergreen-docs.com.
What the app stores
Evergreen AI stores only what it needs to produce and track findings. Page text is processed transiently — held in memory for the duration of a single analysis and the Forge LLMs call — and is not stored in full. Only the following is persisted:
| Data | What it contains | Where it is stored | Retention |
|---|---|---|---|
| Evidence excerpts | Short quoted excerpts from analyzed pages (up to 500 characters each, up to three per finding), the reasoning for a finding, and character offsets. | Forge SQL | For the life of the finding — findings are auto-compacted 90 days after they are resolved (see the row below) — and removed immediately if the finding is deleted or all app data is purged. |
| Findings & triage state | Finding type, confidence score, status, dismiss reasons, and the assignee’s Atlassian account ID. | Forge SQL | Until resolved plus 90 days (automatic compaction), or until purged. |
| Scan ledger | Counts, token usage, durations, and error codes. No page content. | Forge SQL | 12 months, rolling. |
| Page index | Page ID, version, hashed title tokens, and suppression notes used to avoid re-flagging dismissed findings. | Forge SQL | While the page is in an analyzed space. |
| Configuration | Space selections, scan schedules, thresholds, and budget settings. | Forge KV | For the life of the installation. |
| Personal data | Atlassian account IDs of assignees and the people who dismiss findings. Display names are resolved when a screen is rendered and are not stored. | Forge SQL (account IDs only) | As above; subject to deletion as described below. |
Evidence excerpts are quoted verbatim from the pages you ask the app to analyze, so an excerpt may incidentally include personal data that already appears on a page — for example, a person’s name in a “contact” or escalation reference. The app does not seek out, index, or build a profile from such information; it is held as page content under the retention and deletion rules above and is removed when the finding is deleted or when all app data is purged. Beyond the Atlassian account IDs noted above, the app does not store user-profile attributes — such as names, email addresses, or job titles — from the Atlassian user directory.
Where it is stored, and data residency
All app data is stored in Atlassian-hosted Forge storage (Forge SQL and Forge KV), and no app data is stored outside Atlassian infrastructure. Data-residency behavior follows Atlassian’s Forge hosted-storage capabilities and the data-residency status shown for this app in Atlassian admin and on the Marketplace listing.
How AI processing works
Evergreen AI analyzes page content using large language models hosted by Atlassian (Forge LLMs). Page content sent to Forge LLMs is processed by Atlassian-hosted models inside Atlassian infrastructure; it is not sent to any vendor-operated or non-Atlassian service by this app, and it is not used by us (the vendor) for any purpose other than producing the analysis result. No content leaves the Atlassian boundary. We make no representation about Atlassian’s own platform processing beyond Atlassian’s published terms. AI output is advisory: every AI-generated finding includes quoted evidence and is subject to human confirmation or dismissal, and the app never modifies page content. Administrators control which spaces are analyzed; no analysis occurs until a space is explicitly selected.
Subprocessors
Vendor-appointed subprocessors: none. Atlassian provides the Forge platform services (including AI inference) used by the app; Atlassian’s own subprocessor disclosures apply to that platform processing.
Deleting your data
- Admin purge. An administrator can purge all of the app’s data from the app’s settings. This is immediate and irreversible.
- Uninstall. Uninstalling the app triggers Atlassian’s standard Forge hosted-storage lifecycle: data is soft-deleted and then deleted according to Atlassian’s hosted-storage retention and disposal process. The administrator purge inside the app is the immediate deletion route for app-owned data.
- Personal-data requests. Erasure requests keyed to an Atlassian account ID are honored through an account-ID-keyed deletion process, alongside the standard Atlassian personal-data erasure obligations. Personal data that happens to appear inside a stored page excerpt is removed when the related finding is deleted or when all app data is purged.
This website
This website is static. It sets no cookies, runs no analytics, and loads no third-party resources. It does not collect information about visitors. The contact links open your own email client.
Changes to this policy
If this policy changes, the “last updated” date above will change with it. Material changes affecting how the app handles data will also be reflected in the app’s Marketplace listing.
Contact
Questions or requests: support@evergreen-docs.com. See also Security & data handling and Support.