FLINT MARKDOWN EXPORT · CONFLUENCE CLOUD
Export Confluence pages to Markdown.
Choose the pages you need. Preview the files and warnings. Take a Markdown copy into Obsidian, Git documentation, or your next workflow.
- Choose your scope. Select pages from a page tree or space.
- Start with your destination. Use presets for Markdown, Obsidian, Git documentation, or AI context.
- Review what comes with you. Inspect page links, images, attachments and warnings before downloading.
Choose a page, a tree, or a space
- After installing the app, open a page’s more-actions menu, then Apps → Export to Markdown. The Confluence Apps entry opens space exports.
- Choose This page only, This page and descendants, or Pages from a space. For a space, choose it from the Space list.
- Search page titles and use the checkboxes to choose what to export. Branch checkboxes include their descendants, including pages hidden by search. Select all matches adds only matching page titles; Selected only keeps parent context visible.
- Clicking a title inspects that page without changing your selection. The footer shows the number of selected pages.
Only pages readable by the person exporting are eligible. Space exports collect current pages; folder objects provide hierarchy rather than exported page content.
Start with your destination
| Preset | Links | Filenames | Panels |
|---|---|---|---|
| Standard Markdown | Relative Markdown | Title | Callouts |
| Obsidian vault | Wikilinks | Title + page ID | Callouts |
| Git documentation | Relative Markdown | Page ID | Blockquotes |
| AI context | Relative Markdown | Title + page ID | Plain text |
These are starting points. Advanced settings lets you change links, panels, filenames and YAML metadata. Filename tokens are {title}, {id} and {date}, using the page’s last-updated date. ID filenames stay stable after a title change; moving a page can still change its directory.
Frontmatter is on by default and can be disabled. Choose which fields to keep, or add custom key=value metadata. The Git preset omits the export timestamp from page frontmatter. Choosing another preset replaces customized output settings; custom metadata text is kept.
Images are always included when accessible and within the size limits. Enable Include other attachments to add PDFs, spreadsheets and other file types.
Inspect, review, then download
Prepare page preview prepares the inspected page alone, including its images and included files. It does not change your selection or establish the final cross-page links and ZIP paths.
Review selection → Prepare full preview prepares the selected content and attachments with your settings. Switch between rendered output and Markdown source, inspect files and warnings, then choose Build ZIP from preview to package those prepared files.
To skip the preview, choose Export ZIP. Preparation shows progress and supports cancellation. When ready, review the summary and any warnings, then choose Download ZIP. Returning to selection keeps your pages, search, filters and settings.
The preview is a convenience renderer. Your Markdown editor or destination may render some details differently. Obsidian output includes vault links and navigation hints for links in complex HTML tables; inspect the exported vault in Obsidian before adopting it as a long-term copy.
Run larger work in the background
Choose In background beside the export button, then Start background export. Processing runs on Atlassian infrastructure as the requesting user. Keep the job ID to return through History → Find background export, refresh status, cancel an unfinished job, or download a ready ZIP.
Job access and ZIP availability expire one hour after creation. Downloads check the requester, job state and current visibility of exported pages and pages inlined by include macros. Rendered macros can reference additional sources beyond those checks. An already downloaded file cannot be revoked.
The OAuth export API is under development and is not a public production interface yet.
History and reusable settings
History shows your exports with page titles and IDs, space, exporter, settings, counts, bytes, warnings and source. Search checks the loaded entries; use Load more for older entries. Browser records are labeled browser-reported. History is activity metadata, not a complete security audit log.
Confluence administrators can inspect all users’ history and explicitly clear it. Users can delete their own entries. Entries become unavailable after 90 days and are removed by scheduled cleanup. Older records may lack details, and long titles can be shortened in the history display.
Saved profiles keeps up to 20 named sets of output settings per user, including custom metadata values. Save, reuse, replace or delete them. Profiles remain until deleted and do not save the page selection.
Compare with an earlier export
Load its manifest.json under Compare with a previous export to compare page IDs, versions and paths. The comparison reports added, changed, moved and missing pages. A missing page may reflect scope or permission changes; it does not establish deletion.
What is in the ZIP
- One Markdown file per selected page. Child pages retain their hierarchy. Unselected parent pages can remain as directories.
- Optional YAML frontmatter. Available fields include title, page ID, source URL, space key, parent, breadcrumb, labels, author/editor, dates, version and export time.
- Local assets. Accessible images and opted-in attachments sit in a
.assetsfolder beside their page. - Useful links. Links between exported pages use your chosen Markdown or Obsidian style. Links to pages outside the export keep pointing to Confluence.
manifest.json. Page paths, metadata, output options, warnings, skipped items and size-limit omissions.all-pages.md. The selected pages combined in tree order, with rewritten local links.
Formatting and macros
Headings, lists, completed tasks, code, quotes and simple tables become Markdown. Merged cells and complex tables retain HTML structure. Mentions, dates and statuses become readable text. Panels follow the selected style.
Tables of contents, includes and excerpt-includes have dedicated handling. Other macros use Confluence-rendered output where available. A macro that cannot be rendered retains a marker and a warning. Jira issue keys become links; JQL issue lists are currently reported as unsupported.
Limits and recovery
- 500 pages per collection. A larger tree or space must be narrowed before collecting pages. This is a safeguard, not a capacity guarantee for every content mix.
- 75 MiB per file; 300 MiB of attachments per export. Oversized files are omitted and recorded in the manifest, with Confluence links retained. One MiB is 1,048,576 bytes.
- Device and content matter. Browser exports need enough memory for the prepared files and ZIP. Background jobs have a 15-minute execution limit.
- Unsupported content is visible. Whiteboards, databases and other non-page content are reported as skipped. Read failures and conversion omissions remain in the result rather than being silently counted as exported pages.
- Invalid settings stop preparation. Return to selection to correct them or choose a preset.
- An active app license is required. A missing or inactive license shows a notice; contact your site administrator.
Where your data goes
Browser exports prepare content on your device. Background exports process content in Atlassian-hosted Forge functions and temporarily store the ZIP in Forge Object Store. Forge storage also holds activity history and saved profiles. There is no external processing backend or AI provider.
Downloaded exports are ordinary files. They do not inherit Confluence access controls; manage sharing, retention and deletion according to your organization’s policies. Read the privacy policy or contact support.