Migrate from Confluence to XWiki

A planning guide to moving Confluence content, permissions and users to self-hosted XWiki, covering the paid Migrator Pro, the open-source XML filter, what is lost and how the XML export end of life affects timing.

On this page

Migrate from Confluence to XWiki: what the move involves

Most teams start this move because of a renewal, a licensing change, a residency requirement or lock-in to one vendor's cloud. XWiki is the most complete open-source match for Confluence spaces and permissions, but the migration is a planned project, not a file copy. Treat it as a sequence of decisions: which spaces move, which macros and Marketplace apps have a replacement, who owns each space after cutover, and how users and groups are mapped from your directory. The work is usually done by the internal platform team and the documentation owners, often with an implementation partner for the hosting, the migration run and the hardening. Start by confirming your source edition and backup format with the migration tool, because the XWiki Confluence Migrator Pro reads XML and CSV backups from Confluence Server, Data Center and Cloud. Cloud XML site export reaches end of life on December 1, 2026, so a Cloud customer who plans to use XML should check that date against its own schedule. Ask the migrator vendor which artifacts it preserves in a trial run.

Plan the risk around what the tool does not keep. Confluence Migrator Pro keeps page history and migrates attachments, permissions, users and groups, but it does not preserve attachment history, so decide whether old attachment versions must be archived separately. Run a trial migration on a representative set of spaces, compare permissions and page trees against the source, and check the trial limit of 30 pages per migration before you size a full run. Ask any quote to state what is included in the migration license, what the trial limits are, and whether the work covers testing, cutover and post-migration review.

Which Confluence capabilities map to XWiki cleanly, need rework or have no equivalent

This table compares common Confluence capabilities with their XWiki equivalents and shows what the Confluence Migrator Pro carries over. Read each status as a planning signal: clean items need a spot check after import, rework items need decisions before cutover, and gaps need an owner or an archive plan. The migrator keeps page history but not attachment history, so teams that rely on older attachment versions should plan for them separately. Confirm every macro and Marketplace app against your own inventory.

CapabilityXWiki equivalentWhat the migrator doesStatus
SpacesXWiki wiki spacesImports spaces from XML or CSV backupsClean
Page treesNested pages within spacesImports pages with their parent structure; check deep trees after importClean
Users and groupsXWiki users and groupsImports accounts and groups; sign-in must be reconfigured separatelyClean
PermissionsXWiki rights on spaces and pagesImports permissions; map and test restrictions per space with sample usersNeeds rework
AttachmentsPage attachmentsImports attachment files with their pagesClean
Page historyPage version historyKeeps page historyClean
Attachment historyNo equivalent after migrationDoes not preserve older attachment versions; archive them separately if neededGap
Blog postsXWiki blog applicationImports blog posts; confirm the blog application is installed and configuredNeeds rework
MacrosXWiki macros and wiki syntaxConfluence macros need mapping to XWiki macros; confirm each one in your inventoryNeeds rework
Marketplace appsXWiki extensions where one existsDoes not transfer app data automatically; confirm each app against XWiki optionsGap

Which exports and integrations XWiki's migration path reads

XWiki's Confluence path starts with the export you can actually produce, then rebuilds identity and integrations on the XWiki side. Each card below is a decision or check that determines what arrives in XWiki and what has to be rebuilt by hand.

Confluence Cloud space and site exports

Cloud space admins can export a space, and site admins can create a full site backup from the backup manager. XML site and space export reaches end of life on December 1, 2026, and Atlassian recommends CSV after that date. Confirm with the Migrator Pro vendor which of the Cloud export formats its current release reads for your site before you plan around XML. Export the whole site early, because a site backup is the only export that carries the full content and attachment set for a large tenant.

Confluence Server and Data Center backups

The Migrator Pro reads backups from Confluence Server and Data Center as well as Cloud. Server support ended on February 15, 2024, so any Server instance still in use is running without vendor fixes and should be treated as a source with its own risk. Ask the migration team which backup archive layout they expect from your version, and test a single space first to confirm the attachment and user mapping before a full run.

Open-source XML filter stream

The Migrator Pro builds on XWiki's open-source Confluence XML filter stream, so the core conversion logic is available under the LGPL-2.1 license. Teams with in-house developers can read that component and script a conversion without buying the paid extension. The trade-off is that you own the handling of attachments, permissions, users and groups, and the cases the paid extension covers. Confirm the filter version against the XWiki release you deploy, since the component and the extension must match.

Users, groups and permissions

The Migrator Pro migrates users, groups and permissions, but the imported accounts are not a substitute for your identity source. Decide whether XWiki authenticates against your existing directory or keeps local accounts, and map Confluence groups to XWiki groups before the first import. Check how Confluence space and page restrictions translate, because a permission model that looks equivalent on paper can grant access differently once rebuilt. Test with a sample of restricted pages and named users.

SSO connections to rebuild

Confluence sign-in through an identity provider does not carry over as a setting. On the XWiki side, OpenID Connect sign-in is typically provided through the community xwiki-contrib OIDC extension, and Active Directory and Azure Active Directory SSO applications are listed among XWiki Pro offerings. Confirm each connector's version, supported provider and license in your environment before cutover, and test login, logout and group claims for a pilot group.

Third-party integrations and macros

Confluence Marketplace apps and Jira, Slack or Google connectors do not transfer with the content. Inventory every app and macro your pages depend on, then decide per item whether XWiki has an equivalent, a community extension or a custom build. XWiki Pro lists Office365, Google Apps and OnlyOffice connectors, but you must verify each one against your own workflows. Pages that embed these integrations will often import as static content that stops updating.

The migration sequence from inventory to decommission

This sequence moves Confluence content to XWiki in controlled stages, so each space is inventoried, proven on a pilot, run in parallel, cut over with users informed, and only then retired. Each step ends with a written artifact that the next step depends on.

  1. Build the inventory register

    List every Confluence space, page tree depth, blog post, attachment, user, group and space permission, then record which macros and Marketplace apps each space relies on. Assign an owner to each space and flag content that has no XWiki equivalent, such as app-specific macros or Marketplace add-ons. The output is an inventory register that the project sponsor and space owners sign off, and it becomes the baseline for every later reconciliation. Ask the vendor or your partner to confirm which Confluence edition and backup format your site produces before you export anything.

  2. Map users, groups and permissions

    Compare the Confluence user and group list with your identity source, and decide which accounts, inactive users and external guests will exist in XWiki. Translate each space permission into an XWiki group or right and document any permission that does not map one to one. The output is a permission mapping sheet approved by the security or IT owner, so that access decisions are made once, in writing, rather than discovered during testing. Confirm with XWiki how the migrator handles users and groups in your specific export before you rely on it.

  3. Run a pilot on a limited set of spaces

    Select a small set of representative spaces, including one with heavy macros, one with restricted permissions and one with many attachments. Run them through the Confluence Migrator Pro trial, which is limited to 30 pages per migration, so split larger spaces into test batches or use a licensed run for the full pilot. Check the trial terms and the file formats it reads against your export before starting. The output is a pilot reconciliation report that compares page counts, attachment counts, permissions and rendered macros with the inventory register, and lists every defect with an owner.

  4. Fix defects and rehearse the full import

    Use the pilot findings to correct the mapping rules, macro replacements and permission translations, then repeat the import on a full non-production copy of the data. Record how long each run takes so that the production window can be planned from measured results instead of estimates. Have space owners review a sample of pages in the rehearsal and sign the results. The output is a rehearsal log and a corrected runbook that names the exact export, import and verification commands used for each space.

  5. Operate both wikis in parallel with a content freeze

    Keep Confluence available for reading while XWiki takes the working content, and announce a defined content freeze window for each space during which edits are made only in the target after the final import. Run a delta import for pages changed since the last export and reconcile counts again. Track any edits made during the parallel period and confirm whether they need manual transfer. The output is a signed freeze record per space and a delta reconciliation report showing that the XWiki copy matches the frozen source.

  6. Cut over with redirects and user communication

    Switch the default knowledge entry points to XWiki, publish a short guide for users covering new URLs, sign-in, where permissions changed and how to request access, and put a read-only notice on Confluence. Set up redirects or a mapping page for frequently linked URLs and check internal links inside XWiki. The output is a signed cutover checklist covering DNS or link changes, communications sent, support contacts named and the rollback decision owner, with the rollback window recorded before cutover starts.

  7. Decommission the Confluence site

    After the agreed stabilization period, take a final full export and store it in your retention system, confirm that the archive can be read, and then revoke licenses, remove integrations and shut down the Confluence site or instance. Confirm legal hold and records retention requirements with your compliance team before deletion. Ask your Atlassian contact to confirm the terms that apply to your contract. The output is a decommission certificate listing the archive location, checksum or verification result, approvals and the date access was removed.

How the December 1, 2026 XML export end of life changes the plan

Atlassian ends support for Confluence Cloud XML site and space export on December 1, 2026. After that date the export features still work, but Atlassian will no longer fix bugs or accept support tickets for them, and it recommends CSV export instead. For a migration to XWiki, this means the export format is a decision to make early, not a detail to settle at cutover. If an export is produced in a format the vendor no longer supports, any defect you hit during the move is your problem to work around, with no vendor fix behind it.

XWiki's Confluence Migrator Pro reads both XML and CSV backups from Confluence Cloud, so the choice is open. Before committing, run a pilot on a representative subset of spaces with each format. Compare what arrives in XWiki for page hierarchy, attachments, user and group mapping, and permissions, and check how each export handles the content types your spaces depend on. Confirm with XWiki SAS which format its current migrator version recommends for Confluence Cloud sources, and confirm what the open-source Confluence XML filter reads if you plan to use it directly.

Do not plan the migration around the XML path simply because it is familiar. Note the requirement that attachment history is not preserved in either route, so the export you pick should be validated against your own retention and audit needs. Record which format produced each pilot result, keep the export files and the export date, and ask whoever produces the final export to document the Confluence site version and the tool used. That record lets you reproduce or explain the migration later, whatever the timeline turns out to be.

How the migration differs by Confluence edition and starting point

Cloud sites offer two export routes. Space admins can export a space as PDF, CSV, HTML or XML, and site admins can create a full site backup from the backup manager. The XWiki migrator reads XML and CSV backups from Cloud, so the export format you choose now shapes the import. You can read Atlassian's export documentation for the steps and the limits of each format.

  • Space exports: Space-level exports give you the most control and suit a phased move, one space at a time.
  • XML exports: Atlassian's XML site and space export reaches end of life on December 1, 2026. Features keep working after that date, but Atlassian will not fix bugs or accept support tickets for them, so an XML-based plan should finish before then.
  • CSV exports: Atlassian recommends CSV after XML ends. Ask XWiki SAS which content types the CSV path carries into XWiki, especially attachments and permissions.
  • Full site backup: This route needs site admin rights and suits a single whole-site move. Confirm with XWiki SAS what the backup contains before you plan the import around it.

Confirm the export format you will use, the date you plan to run it, and whether attachment history matters to your users, because the Migrator Pro keeps page history but not attachment history.

What drives the cost and duration of a Confluence to XWiki migration

What does the Confluence Migrator Pro license cost for an XWiki migration?

XWiki SAS sells Confluence Migrator Pro as a paid extension, and its XWiki Store listing gives a starting price of 2000 EUR. The listing also offers a trial limited to 30 pages per migration, which lets you test mapping on a sample before buying. Confirm with XWiki SAS whether the quoted price covers one migration run or a license for a set period, whether it scales with the number of spaces, pages or attachments, and whether re-runs for cutover or delta imports are included. The license is a one-time purchase for the migration tooling, so it does not recur on its own.

Is the open-source Confluence XML filter a free alternative to the paid migrator?

The open-source XML filter stream is the component that Confluence Migrator Pro builds on, so it is free to use under XWiki's open-source licensing. The trade-off is effort: your team or a partner must script the import, map users and groups, handle attachments and permissions, and verify the results without the paid extension's packaged workflow. Estimate the internal engineering time for that work before comparing it with the license fee. Ask any integrator to state in writing which parts of the migration they will build and which they will test.

How much does XWiki SAS consulting cost per day, and what must we buy first?

XWiki SAS lists consulting at 1200 EUR per day, and that rate requires a Business or Enterprise support plan. The support tier therefore becomes a precondition for this rate, so its recurring fee should be added to the consulting estimate. Ask XWiki SAS to confirm the number of consulting days needed for your spaces, macros, user directory and permission model, and whether the day rate applies to remote work, on-site work or both.

What does XWiki training cost, and when is it worth buying?

XWiki SAS lists training at 1500 EUR per day. Training is one-time for the people who will administer the wiki, but its value depends on how many administrators, content owners and power users you need to onboard. Teams with strong Java, Tomcat and database skills may need less formal training than teams who will build XWiki applications. Ask for a course outline that covers administration, permissions, macros and the migration workflow, and confirm whether training can be delivered on your own infrastructure.

What hosting costs recur after a self-hosted XWiki migration?

Hosting is the main recurring infrastructure cost. It covers compute for the Tomcat application, a database (MySQL, PostgreSQL or MariaDB, depending on the official image variant you choose), storage for attachments, backups, monitoring, TLS certificates and network access. Costs rise with user count, attachment volume, the number of environments (staging and production are usually both needed), and availability targets. Get the hosting assumptions written into the quote, including whether they run in your cloud account, your data center or a provider's infrastructure.

What internal admin time and content cleanup should we budget for?

Internal time is often the largest cost and is easy to underestimate. Your team must decide which Confluence spaces, pages and attachments to keep, assign owners, rebuild or review macros and Marketplace app content, confirm group and permission mapping, test sample pages and train authors. Content cleanup is one-time work, but it recurs in smaller amounts when old content is archived or re-imported. Ask any partner to list which cleanup tasks are in scope and which fall to your staff.

Which costs are one-time and which recur, and what should a written quote include?

One-time costs include the migrator license, migration engineering, content cleanup and initial training. Recurring costs include the XWiki support subscription, hosting, backups, upgrades, security patching and the admin time to run the wiki. A written quote should separate these lines, state the scope in writing, name the Confluence edition and backup formats covered, list the spaces and page counts assumed, explain what is excluded (such as attachment history, which the migrator does not preserve), and give payment and change-request terms. Ask for the upgrade path between XWiki release lines, since stable and LTS versions follow different schedules.

What usually goes wrong and how to prevent it

Confluence to XWiki migrations tend to fail in predictable places: content that does not carry over, identities that do not line up, and pages that change while the copy is being made. Each card below names one of these risks and a check to run before cutover, so the decision to proceed rests on evidence from your own export.

Attachment history does not carry over

XWiki's Confluence Migrator Pro keeps page history but not attachment history, so earlier versions of uploaded files are not carried across. Teams that rely on old contract or policy versions stored as attachments need a different archive. Confirm which attachments need prior versions and export those files separately before cutover.

Macros and Marketplace apps without equivalents

Confluence macros and Marketplace apps do not map one to one onto XWiki. Pages that depend on a specific app may import as static text or broken blocks. Inventory macros across the page tree, mark each one as rebuilt, replaced or dropped, and sign off each gap with the page owner before the parallel run starts.

Permissions that map imperfectly

Confluence space permissions, page restrictions and group roles are expressed differently in XWiki's rights model. Some restrictions may be widened or narrowed during import. Pick a sample of restricted spaces and compare who can view and edit before and after the migration, using test accounts in each group.

User identity mismatches

Users and groups are matched by identifiers in the export, so accounts with changed usernames, duplicate emails or deactivated profiles can end up as orphaned authors or lose access. Reconcile the user list against your directory first, decide how departed users are represented, and check that authorship and group membership on a sample of pages after import.

Large attachment volumes

Large attachment sets increase export size, import time and storage on the new server, and an interrupted import may need a rerun. Measure total attachment size per space before scheduling, confirm the export is complete, and run a test import on a copy of the largest space to observe timing and disk use on your own hardware.

Edits during the parallel run

Content edited in Confluence while XWiki is being loaded creates two versions of the same page that drift apart. Set a content freeze date for each space, or agree a delta process to re-import changed pages, and record which system is the source of truth for every space before users start working in the new wiki.

Broken internal links

Links between pages, spaces and attachments in Confluence use identifiers and paths that may not resolve in XWiki after import. Readers then land on error pages or on the wrong content. Crawl the imported wiki for broken links, compare results against a list of key pages, and fix or redirect the important ones before announcing cutover.

Questions teams ask before migrating from Confluence to XWiki

How much does it cost to migrate from Confluence to XWiki?

The cost has three parts: the XWiki software, the migration tooling, and the people who do the work. XWiki Standard is free and open source under the LGPL 2.1, so there is no licence fee for the core platform. The Confluence Migrator Pro from XWiki SAS is a paid extension, and the XWiki Store lists it from 2000 EUR, with a trial limited to 30 pages per migration. Open-source XML filter use avoids that fee but shifts more work to your team. Ask any quote to separate one-time migration effort from recurring hosting, support subscriptions and admin time.

How long does a Confluence to XWiki migration take?

The duration depends on the number of spaces, pages, attachments and users, how many macros and custom apps need rework, and how many test cycles your business requires. None of those factors has a fixed duration in the sources we rely on, so no single timeline applies. Ask for a timeline tied to a written scope that lists the spaces in wave one, a pilot space, and the acceptance criteria for each cutover step.

What does the migration need from our team?

Your team needs to supply a Confluence export or backup, access to the source site for user and group data, and business owners who can confirm which spaces, permissions and pages must stay. Engineers should be ready to check macros, Marketplace apps and internal links, which rarely convert without review. Plan for an identity contact who can map users and groups to the new directory or sign-in method, and a named owner who signs off on each space.

Can we run the migration ourselves?

Yes, if your team has administrators comfortable with Tomcat, the chosen database, and XWiki's import tooling, and time to test. XWiki publishes official Docker images in MySQL, PostgreSQL and MariaDB variants, which helps a self-managed build. Teams without that experience often bring in an implementation partner for the migration, hardening and backup setup. DeployOpen plans, implements and supports self-hosted open-source platforms in your own infrastructure under a written, fixed scope, and can quote that work separately from any XWiki SAS services you choose.

Does the migration keep Confluence attachment history?

No. The Confluence Migrator Pro keeps page history but does not preserve attachment history, so older versions of attached files are not carried into XWiki according to the XWiki Store. Confirm this against your Confluence edition and export format before cutover. If older attachment versions matter for audit or records, keep a copy of the source backup and the original files in archive storage, and agree in the written scope how those files will be retained and found.

Is the open-source Confluence XML filter enough for our migration?

It can be enough for simple content, but it is a component rather than a finished migration tool. The XWiki Confluence Migrator Pro builds on the open-source Confluence XML filter stream and adds the packaged import workflow, support, and coverage of pages, spaces, attachments, permissions, blog posts, and users and groups. Use the open filter if your team can write and test its mapping, and compare its output against a Migrator Pro trial on a representative space before deciding.

What happens to Confluence after cutover?

Confluence stays in place until you decommission it, so plan the retirement as its own step. Atlassian ends support for Confluence Cloud XML site and space export on December 1, 2026, and recommends CSV export after that date, so any XML backup you rely on should be taken and archived before then. Data Center licences expire and become read-only on March 28, 2029. Confirm with Atlassian which export and retention options apply to your edition, and agree in writing how read-only access, archives and user redirects will be handled.

How DeployOpen scopes and runs a Confluence to XWiki migration

DeployOpen is a Chainscore Labs company that plans, implements, customizes, integrates and hands over self-hosted open-source platforms inside the client's own infrastructure. For a Confluence to XWiki migration, we start with a written scope that fixes which Confluence sources, export formats and content types are in play before any import runs.

How we scope

  • Inventory the Confluence edition, export formats and backup files you actually hold before choosing a route.
  • Agree in writing which spaces, users, groups, permissions and attachments move, and which history is accepted as lost.
  • Confirm whether the paid Confluence Migrator Pro or the open-source XML filter fits your sources and budget.
  • Set the target hosting, database variant and identity integration in the written scope before work begins.

How we build

  • Deploy XWiki in your own cloud account, data center or air-gapped network using the official Docker images.
  • Run test imports on a copy of the export and compare page counts, permissions and attachment links with the source.
  • Map Confluence groups and space permissions to XWiki rights, and review each mapping with your access owners.
  • Document macros and Marketplace apps that have no XWiki equivalent so they can be rebuilt or retired deliberately.

How we hand over

  • Deliver the migration runbook, configuration records and a reconciliation report listing what moved and what did not.
  • Provide an acceptance checklist covering content, permissions, users and search for your business owners to sign off.
  • Train administrators on backup, upgrade and extension management for the self-hosted XWiki instance.
  • Keep a documented rollback path until the cutover is accepted in writing.

How we support

  • Engagements are fixed-scope, with scope, deliverables and timelines agreed in writing before work starts.
  • Ongoing support, upgrades and continuous testing are available as a retainer after handover.
  • Ask any quote to itemize implementation, migration tooling, testing, training and recurring support separately.

Talk to our team.

Tell us what you're working on, whether it's a deployment, an audit, a security test or a cyber range. You'll speak with an engineer who can help you scope it.

  • 30-minute call: free, with no obligation.
  • NDA on request: we can sign before you share details.
  • Clear next steps: a scope and plan after the call.