Notion Backup: What You Can Recover From an Export

A Notion backup has two useful outcomes: keeping information readable outside Notion and recovering a place where your team can work. A folder of notes and spreadsheets can deliver the first while leaving substantial rebuilding for the second.

Imagine a project tracker with a Tasks database linked to Projects. A project page displays its tasks and a rollup of their estimated hours. Recovering the task names is useful. Resuming work also means knowing which tasks belong to which project, making the rollup calculate again, and opening the files attached to those tasks.

Choose an export format for your recovery goal

HTML exports provide pages you can browse, with assets in accompanying folders. Markdown & CSV provides text files for ordinary pages and CSV tables for full-page databases, with Markdown files for subpages. Keep the surrounding folders with the documents.

HTML can include page-level and block-level comments, including resolved comments. That does not mean those discussions become native comments when you bring the archive back: Notion’s HTML importer excludes comments and edit history. Preserving a readable record and restoring its collaboration features have different requirements.

For workspace-wide backups, use HTML or Markdown & CSV. Notion’s workspace PDF retirement notice gives August 31, 2026 as the end of its rollout; individual-page PDF exports remain available.

Download a workspace backup

On desktop or web, open Settings → General → Export all workspace content. Choose the format and content options offered for the export. Download the archive from Notion’s email: the download link expires after seven days. Keep the downloaded ZIP and an extracted working copy in a location you can reach independently of Notion.

The export follows the exporter’s access. Other users’ inaccessible private pages are omitted, and teamspace settings can exclude content. Before relying on the archive, compare its coverage with the projects you intended to preserve. Notion also explicitly says that uploading an export cannot instantly recreate a workspace. These limits are in its backup instructions.

For a smaller backup, open a top-level page’s export menu and include its subpages. Select the content option that retains the files and images you need. If you collect information through a database form, export its questions and responses from Table view; Form view itself is not exportable.

Label the copy with the export date, workspace and scope. For a team archive, record the exporting account as well. If a client project is absent later, that record helps distinguish a page outside the export’s permissions from a document lost during a subsequent move.

Recover the database relationships

A CSV gives you rows and columns. A Notion relation connects a row to another database’s page. On export, relation properties become plain-text URLs; importing that CSV does not automatically re-establish those original connections.

In the example tracker, a task can retain a reference to its old project page while still lacking a functioning relation to the project you are rebuilding. Preserve those references as clues, then match the recovered tasks to the intended projects. Check the matching on several tasks before rebuilding calculations that depend on it.

The CSV import rules add two practical constraints. A new CSV import cannot create new relations, formulas or rollups. Importing into an existing database can map a column to an existing relation property, but that option is not a promise to reconstruct an exported workspace. CSV imports and merges also add rows rather than updating existing ones.

Use a separate recovery area for the first attempt. Merging yesterday’s task CSV directly into today’s live tracker can add duplicate tasks. In the example, compare task titles and project assignments in the recovery copy, recreate the required relation and calculations, and check the resulting project total before deciding what belongs in the working tracker.

For this sample tracker, use Notion’s relation and rollup controls to rebuild the connection:

1. Import Projects and Tasks into the recovery area.
2. Add a Relation property to Tasks, select the recovered Projects database, and enable the corresponding relation in Projects.
3. Assign each sample task to its recovered project.
4. In Projects, add a Rollup using that task relation, the numeric Hours property, and Sum.

Open each recovered task as a page, too. If its working notes or attachments are missing, use the saved Markdown and files to fill the gaps in the recovery copy. Keep track of that manual work when estimating how long the full tracker would take to recover.

Rehearse with one small project

Choose a project containing the kinds of information you would struggle to replace: a nested brief, several linked tasks, a calculation, an uploaded document and a relevant comment thread. Write down what you expect to recover from each. This gives the rehearsal a useful finish line beyond “the import completed.”

Create a separate private recovery area. For the first pass, import only the sample’s notes and CSVs through Settings → Import. Keep a simple inventory of the original sample so you can compare it with the recovered copy.

For a linked external document, also verify recovery through its original service. Opening a link while signed into your usual account proves access at that moment; it does not demonstrate that the Notion archive contains the document itself.

Notion’s Share menu shows who can access a page and their permission levels. Inspect the destination’s audience before moving recovered client material into a teamspace. Pages in a default teamspace are accessible to everyone in the workspace except restricted members.

Save the reconstruction steps alongside the archive. A useful record identifies what imported cleanly, what you rebuilt, and what remains unavailable. Repeat the sample when a tracker gains a new database or a new kind of attachment.

Use Trash and version history for recent mistakes

For an accidentally deleted page, check Trash before importing an archive. Pages remain there for 30 days by default; Enterprise owners can customize retention settings. For an overwritten page, open its ••• → Version history. Notion’s recovery guide lists these history windows:

You need at least Can edit access to view page history. Copy specific blocks from an earlier version when you only need a missing section; restoring the whole version makes it current.

Database recovery has an important boundary. Restoring a database restores its pages and properties, but not the contents inside every database page. Recover those bodies from each page’s own history. If a task’s status and its written handover were both damaged, check both after recovery.

Compare automatic backups by their restore process

Notion’s backup FAQ says it does not currently offer scheduled automatic exports. An external service can handle a schedule, but the useful comparison is what happens when you need the copy back.

For example, BackupLABS’ Notion page describes recovery by downloading a ZIP and copying or pasting data back into Notion. Notion Backups’ FAQ describes restores to the same workspace or, on a higher plan, another workspace. It also acknowledges that API limitations prevent a complete restore. These are the vendors’ descriptions of their services.

Before subscribing, ask the provider to walk through your sample project. Does recovery create new pages or change existing ones? What happens to relations and calculations? Which comments and attachments come back, and where? Ask for the unsupported-content list and how failed or partial backups are reported.

Also establish where the retained files live and how you can retrieve them if the subscription ends. A schedule is valuable when it saves you repeated work; the restore process determines how much work remains after a loss.

What an API backup must collect

If you build your own backup, retrieving a page object is only the beginning. Notion’s page endpoint returns properties; the page body requires content retrieval. Large relation properties can also need additional requests, and the connection must have read access.

The block-children endpoint is paginated and returns one level at a time. A collector must follow the remaining pages and descend into nested blocks. Record missing access and failed requests so that an apparently successful run does not conceal an incomplete project.

Download attachments while their links are valid. Notion-hosted file URLs expire after one hour; storing the URL alone leaves you with an expired reference. Preserve the file contents and their association with the relevant page. External URLs need their own availability check.

Build retry handling around the current API limits. Connection budgets depend on the workspace plan, with additional workspace and endpoint limits. A 429 response supplies a Retry-After value. Log what the job actually collected, then use the same small-project rehearsal to confirm what your recovery process can rebuild.