Changelog - June 2026 #1
What's New
- Restore encrypted Notion backups: Encrypted Notion backups can now be restored end to end. Your archive is decrypted at restore time using your private key, which is never stored and is discarded immediately after use.
- Back up your Notion people directory (optional, off by default): Capture your workspace's member directory as an access audit trail and to help resolve "assigned to" and "created by" fields. Note this is backup-only, since Notion's API cannot recreate users on restore.
Notion Backups
- Formulas and relations survive a restore: Formula properties no longer come back empty, and relations to databases that were also backed up now relink correctly instead of silently dropping.
- A clearer restore page: Removed a redundant "choose your restore method" step, added a plain-language note that restoring never overwrites your existing workspace, and turned the confusing empty state into an actual checklist with copy and refresh actions.
- Handles unsupported blocks gracefully: A single unsupported block (such as an AI block) no longer causes an entire page of content to be dropped. Everything around it is preserved.
- Comment backups fixed: Notion comment backups were silently failing. They now capture correctly, and the backup counters no longer report "0 backed up" on runs that actually succeeded.
- Handles complex workspaces: Workspaces with circular page and database references no longer cause a backup to crash. Cycles are detected and skipped mid-run.
- Clearer backup logs: The Content tab now uses plain labels, groups failures by whether they will auto-retry or need your attention, shows humanized messages instead of raw errors, and separates issues on our end from ones you need to act on.
- Steadier long runs: Better handling of long Notion backups jobs, allowing better logging of large Workspace backups.
- Smoother OAuth: Connecting Notion no longer leaves you stranded on a dead tab when Notion hands off to its desktop app, which also resolves duplicate connections.
Databases
- One file per database restored: MySQL all-databases backups again produce one file per database (matching previous behaviour), so you can restore a single database instead of the whole set. Backups from the affected window are migrated automatically.
- Managed MySQL support: On managed providers like DigitalOcean, all-databases backups now fall back gracefully when the database user cannot read system schemas, instead of failing.
- Postgres in search: PostgreSQL now appears when you search for it in the backup source picker.
- Post-command scripts always run: Scripts that should run after a backup now run even when the backup fails partway through, so steps like resuming Postgres WAL replay aren't left hanging.
Backups & Reliability
- Clearer snapshot errors: Cloud provider snapshot failures now show the actual reason from Linode, Vultr, or Scaleway instead of a generic code.
- Stale locks release automatically: Improved locks release mechanisms.
- Reliable Scaleway listings: Snapshot, volume, and database listings improved.
- Google Cloud Storage guidance: Added inline guidance for GCS buckets added through the S3-compatible provider, pointing you to Fine-grained access or the native GCS provider.
- More accurate anomaly notifications: Small backup folders no longer show as "0 B" in anomaly notifications.
Billing & Interface
- Payment method fixes: Card-free payment methods such as Link and SEPA no longer make a successful setup look like a failure.
- Clearer upgrades: The upgrade dialog now highlights which plan to pick instead of leaving it ambiguous.
- Team Reports permissions: Reports are now restricted to admins and owners.




























































