Exporting your data
Last updated
Download a copy
In Settings → General, choose Export data. Your browser downloads a ZIP of UTF-8 JSON files. Only the signed-in owner's workspace is included, including archived items. Keep the archive private: it contains your work and answers.
Downloads stream directly to your browser. Allow one to finish before starting another; you can start up to three per hour. Cancelling the download stops the export. If a download fails or your ZIP tool reports an incomplete archive, start a new download. The app does not store a prepared archive for later.
What is included
The archive contains brands, projects, batches, tickets and their spec history, criteria and relations; runs, events, findings, artifacts and requests (questions, approvals and decisions); dependency proposals, submissions, human reviews and AI review advice; playbooks and their versions, ticket templates, recurring schedules and occurrence history; worker details and project access; notification preferences, reminder attempts and undelivered-alert records. Format v2 also includes workspace membership records: user ID, role, status, creation/change times, changed-by ID, and the co-member's display name and email. Only memberships of the exporting workspace are included, including suspended/removed history; a shared person's memberships in other workspaces are excluded.
Worker credentials, authentication/session data and private notification topics or access tokens are excluded. Text uses the same secret masking as the app; redacted content stays redacted. No other workspace's data is included.
Read the format
manifest.json identifies format getworkers-export, version 2. Its files
map describes every data file and field type. Each data file is a JSON array;
an empty dataset is []. IDs join records between files, including references
to archived items. Optional deleted/omitted references may no longer resolve.
complete.json records when the stream finished. A valid ZIP and that file
indicate a finished download.
Dates are ISO 8601 UTC strings. Decimal costs are strings; unknown costs remain
null, never zero. Enum values use the stored uppercase spelling. The schema
uses ? for nullable values and [] for arrays. Json fields contain nested
JSON, with credential-shaped properties removed and text masked. Ticket spec
snapshots preserve their original versioned shape. Future incompatible formats
will use a new format version; keep the manifest with the files.
Reading an older archive
Version 1 archives remain readable: open their own manifest and follow its embedded file/field map. Both versions use the same ZIP, JSON arrays, IDs, UTC dates, decimal strings, null values and completion marker. Keep the manifest with the files; a newer schema must not be applied blindly to an older archive.
Version 2 adds memberships.json, organisationId on projects, tickets, workers
and events, and sponsorMemberId on workers. That sponsor ID points to a
membership record. These fields describe the archived workspace and history;
they are not permission to access anything in the running app.
In version 1 those additions were not recorded. Missing membership or sponsor
data is unknown, not an empty list or a role you can infer from ownerId.
Use manifest.organisationId and existing parent IDs to read the old hierarchy;
v1 cannot recover membership history it never included. This feature does not
import, restore or upgrade old archives.
Consistency and limits
This is a live export, not an atomic database snapshot. Pause edits and worker activity while downloading if you need a consistent copy. Changes made during the download can appear at different points in different files. Recheck the archive after a quiet export before using it to move elsewhere.
The format is for reading and portability. This feature does not import or restore data and cannot restore excluded credentials. Link artifacts contain their recorded URL, not the contents of external files. The export does not include public beta sign-ups or internal transport/idempotency bookkeeping.