From sign-up to a fully encrypted, restorable snapshot of your Confluence Cloud site - spaces, page trees, blog posts, folders, comments, attachments, labels, properties, restrictions, space permissions, users and groups. This guide walks you through every step.
Fill in your work email, company and a password. Click Start free trial.
Confirm via the verification email.
Your trial workspace is provisioned instantly - every application, Confluence included, is available during the trial.
Painless onboarding
Trial accounts get the full Confluence feature set - AES-256 encryption, BYOB, scheduling, granular restore, recycle bin. Switch to a paid plan at any moment without losing data.
Option B - Paid licence (Stripe checkout)
Go to cybback.com/tarifs and pick the tier that matches your storage needs. CYBBACK is priced per protected user: a Confluence user is an active Atlassian account with Confluence access (bots and deactivated accounts are never counted).
Click Subscribe. Stripe Checkout supports cards, SEPA and invoicing on annual plans.
Your licence is active immediately - visible under Account → Subscription.
2
Connect Confluence
Create the service account & token
CYBBACK talks to Confluence through the Confluence Cloud REST API, authenticated with an Atlassian API token. No Marketplace app to install, no OAuth consent screen.
Recommended - an Atlassian service account
In admin.atlassian.com → Directory → Service accounts, create a service account (e.g. Backup). A service account does not consume a Confluence licence.
Under Apps, give it the Confluence → User role and the Confluence → App admin role. App admin lets it see every space, including spaces it is not a member of.
Create an API token for the service account and select the scopes listed in CYBBACK's Quick setup guide (Configuration tab): the read scopes for the backup and the write scopes for the restore. A Copy all button copies each list.
If you want groups and their members restored, the account also needs User access admin on the organisation.
Already protecting Jira?
The same service account can serve Jira and Confluence. Create one token per module - an Atlassian token holds at most 50 scopes.
Alternative - a dedicated user with a classic token
Invite a user such as backup@yourcompany.com with Confluence access and the Confluence administrator role (this consumes a licence).
To restore into a space that still exists, the account needs that space's permissions: add / edit / delete pages and blogs (Edit is required to overwrite a page that still exists), add / delete comments and attachments, and Admin for restrictions and space permissions. The simplest setup is to make it a space admin. Spaces recreated by a restore belong to the account.
Enter your site URL (https://yourcompany.atlassian.net), the account email and the API token, then click Validate. CYBBACK identifies the account, lists your spaces and checks the restore permissions. A scoped token is detected automatically and routed through api.atlassian.com.
The token is encrypted (AES-256-GCM) and stored in the CYBBACK vault. It is never shown in clear again.
A missing scope is not a broken token
If a scope is missing, only the matching data is skipped and the backup reports a clear warning naming it - the rest of the site is still backed up. Add the scope to the token and the next run picks it up. A revoked token, on the other hand, is flagged immediately (no silent retries).
Two common oversights. Without the Confluence User and App admin roles, the account only sees public spaces: validation says so (“no Confluence app role”). Without the readonly:content.attachment:confluence scope, attachments are not downloaded; without read:template:confluence, global templates are skipped. A token cannot be edited: create a new one with the full list and save it in CYBBACK.
View-restricted content. A page, whiteboard or database whose view access is limited to specific people is invisible to the backup account, even as App admin: add the account (or one of its groups) to those restrictions so they are backed up. Blog posts. They can only be restored into a space where the blog is enabled; otherwise they are skipped and the restore says so. With Confluence Premium or Enterprise, your CYBBACK contact can enable reading such content through the Confluence admin key, without changing its restrictions - Atlassian only accepts it with the classic API token of an organisation or site administrator, not with a service account. Databases. Their fields, views and entries are backed up and exported (JSON, CSV); Confluence cannot write them back, so a restored database is recreated empty. Whiteboards. Title and position only: their content is not exposed by Confluence.
Pages and blog posts - body, parent and position in the page tree, status (current or archived).
Folders, plus whiteboards and databases as metadata only (their content is not exposed by the public API).
Labels, content properties and restrictions of every page and blog post.
Version history metadata - number, author, date and message of every version.
Deleted content - anything that vanished since the previous run is kept in the CYBBACK recycle bin, even after the Confluence trash has been emptied.
Options (toggles)
Attachments - the files and images themselves, downloaded into your bucket (on by default).
Comments - footer comments with replies, inline comments with the highlighted text (on by default).
Previous versions - the bodies of earlier versions (off by default, costly on large sites).
Personal spaces, archived spaces, users & groups - on by default.
Dashboard → Confluence → Configuration
Backup scopeSite: acme.atlassian.net
ENG - Engineeringglobal
HR - People handbookknowledge base
~jdoe - John Doepersonal
AttachmentsFiles and images of pages and blog posts
CommentsFooter and inline comments
Incremental forever
After the first full backup, CYBBACK only fetches content modified since the previous run (CQL lastmodified, space by space). Unchanged pages keep their comments, labels and attachments from the previous run without any API call. A daily backup of a large site transfers virtually only the delta.
4
Bring Your Own Bucket
Use your own S3-compatible storage
Want full data sovereignty? Point CYBBACK at your S3-compatible bucket - AWS S3, GCS, Scaleway, OVH, Wasabi, MinIO. Your data, your provider, your region.
Provision the bucket
In your S3 provider, create a private bucket (no public access, versioning recommended).
Create an access key / secret key with permissions limited to that bucket.
Required IAM actions:PutObject, GetObject, DeleteObject, ListBucket.
Resilience: CYBBACK stores an encrypted snapshot of S3 credentials with each backup. Old backups remain restorable even after you switch providers.
Optional encryption at rest: enable AES-256-GCM encryption so every JSON file (.x) and attachment (.xb) is encrypted before upload.
5
First backup
Run your first backup
You're now fully configured. Time to run your first Confluence snapshot - manually for instant peace of mind; a daily schedule is created automatically with it.
Manual backup
Go back to the Dashboard tab.
Click Run backup now. The job appears with status pending, then running, with a progress bar and the current phase (site, users & groups, spaces, contents, attachments).
Once finished, the history line shows the counts (spaces, pages, blog posts, comments, attachments) and the size really stored by this run. Partial failures and skipped scopes are listed under the line.
Automatic daily schedule
The first backup creates a schedule Auto - Confluence (daily at 03:00 Europe/Paris) - a toast confirms it, and the dashboard shows the next run.
Real-time: the Events page streams every operation across all services via SSE.
Notifications: success / failure / drift alerts via email, Slack and webhooks.
Drift detection (opt-in): alerts on mass deletions of spaces, pages, blog posts or attachments between two runs.
Atlassian rate limits
Confluence Cloud applies cost-based rate limits per site and answers 429 when exceeded. CYBBACK applies an internal token bucket, honours Retry-After and backs off automatically - so a very large site simply takes a little longer on its first full run.
6
Restore
Restore - page, page tree, space
CYBBACK restores from any previous snapshot. Recreate a single page, a page with all its children, a blog post, a whole space, or recover deleted content from the recycle bin - into the original space or a different one.
The 3-step restore flow
Selection. Open the snapshot, browse the spaces and the page tree, search by title, preview any page and tick what you want to recover. Toggle Recycle bin to see content deleted from Confluence since the backup.
Options. Choose your safety net:
Dry-run - simulate, no write to the Confluence API. The worker lists what it would create.
Target space - leave empty to restore into the original space (recreated if missing), or pick another existing space.
Include children - restore the selected page with its whole sub-tree.
Comments, attachments, labels, properties, restrictions, space permissions, history - each can be included or skipped.
Execution. The async worker recreates parents before children in their original order, uploads the attachments, re-embeds the images in the page body, then reposts comments and applies labels, properties and restrictions. Live logs and counters are shown on the page; a summary lists what was created, reused, skipped or failed. Re-running the same restore does not create duplicates.
Dashboard → Confluence → Restore
1. Selection
2. Options
3. Execution
Choose itemsSnapshot from Nov 10, 2026
Space - ENG Engineering120 contents
Architecture decision recordsdeleted · recycle bin
Release checklist4 attachments
Onboarding12 child pages
Space - HR People handbook48 contents
Best practices
Always start with a dry-run - the worker reports exactly what it would create without touching Confluence.
Restore to a test space first for major DR scenarios. Validate the result, then re-run to the final target.
Titles are unique per space - a title already taken in the target space gets a (restored) suffix.
Honest restore limits
The Confluence API recreates content under the restore account: authors and creation dates of pages, comments and versions cannot be restored (the original author and date are annotated on each comment). Version history is exported and can be attached as history.json, but it is not replayed. Inline comments cannot be re-anchored through the API: they come back as footer comments quoting the highlighted text. Whiteboards and databases are recreated empty, because their content is not exposed by the API. A restriction that cannot be applied leaves the page restricted to the restore account - never more open than the original.
You're all set.
Need help? Our team replies within one business day on every plan.