BlueConfetti

FLINT MARKDOWN EXPORT · HOW-TO

Break a large Confluence space into manageable exports

When a space is too large for one export, split it into branches you can track, repeat and reconcile.

Finished result

A batch plan with one export per branch, a manifest for each batch, and a duplicate-page check and a way to investigate gaps while the space changes.

Why one export is not enough

Flint collects up to 500 pages for one export. The cap applies while the page list is being collected, before you can choose pages. When a tree or space is larger, collection stops and the workspace shows Could not complete the request with this message:

This section has more than 500 pages; the limit is 500. Export a smaller section.

Narrowing the selection afterwards is not possible, because the list never loads. Start from a smaller branch instead. The other limits are per export too: 75 MiB per file and 300 MiB of attachments. Background exports share these limits and also have a 15-minute execution limit, so running in the background does not lift the cap.

Plan the batches

  1. Use the space’s page tree to list its top-level branches. Each branch root becomes the starting page of a batch.
  2. Export the space home page on its own with This page only; exporting it with descendants collects its tree and can exceed the cap.
  3. If a branch stops with the page cap message, split it into its child branches.
  4. Keep attachments in mind. A branch with many large files may need splitting even when it has fewer than 500 pages.
  5. Use one saved profile for every batch so the files are consistent; see export profiles.
BatchStart pageSource optionNotes
HomeCustomer Support (7000)This page onlySpace home page only
AProduct guides (7100)This page and descendants7 pages
BTroubleshooting (7200)This page and descendants7 pages after a move, see below
CRelease notes (7300)This page and descendantsExport last

The plan is illustrative. For each finished batch, record the archive name, the root page ID and page count from manifest.json, and the export time.

Run each batch

  1. Open the batch’s start page and choose Apps → Export to Markdown.
  2. Choose the source option from the plan and keep the collected pages selected.
  3. Open Saved profiles, choose Use profile, and set Include other attachments, which a profile does not store.
  4. Export, then download the ZIP and keep its manifest with the batch record.

Check for overlaps and gaps

A space can change while you work through the batches. In the sample, Batch A was exported at 16:00. At 16:10 an editor moved Product guide 1 (7101) into Troubleshooting, so Batch B, exported at 16:30, contains it again. These commands were tested with jq 1.7.1 on the sample manifests:

jq -r '"\(input_filename): root \(.root_page_id), \(.pages | length) pages, exported \(.exported)"' batch-a-manifest.json batch-b-manifest.json
jq -r '.pages[].page_id' batch-a-manifest.json batch-b-manifest.json | sort | uniq -d
jq -r --arg id 7101 '.pages[] | select(.page_id == $id) | "\(input_filename): \(.path)"' batch-a-manifest.json batch-b-manifest.json
batch-a-manifest.json: root 7100, 7 pages, exported 2026-11-03T16:00:00Z
batch-b-manifest.json: root 7200, 7 pages, exported 2026-11-03T16:30:00Z
7101
batch-a-manifest.json: 7100/7101.md
batch-b-manifest.json: 7200/7101.md

The duplicate is easy to spot. The opposite case is harder: a page moved from a branch you have not exported yet into one you already finished appears in neither batch. Count the pages you expect per branch before you start, compare each manifest with that count, and re-export any branch that changed during the run.

Sample files

Limits

Sample files on this page are fictional. They were produced by the released Flint Markdown Export exporter, using a simulated Confluence site instead of a live one. Interface labels come from the released app; this guide contains no app screenshots.