Import and export

The public CLI provides dataset and asset transfer. Use the import, export, and assets references for exact options.

Validate a round trip

Confirm the source project and local schema configuration.
Export to a new filename and inspect the archive contents.
Prepare a disposable destination project with the matching schema and scopes.
Import the archive into that destination with an explicit project selection.
Compare record counts and IDs, relationships, scoped values, asset metadata, and downloaded files.
Re-run a small controlled import to verify the chosen skip/update behavior before relying on repeated imports.

flexkit export catalog.tar.gz --project SOURCE_PROJECT_ID
flexkit import catalog.tar.gz --project DESTINATION_PROJECT_ID

Replace both IDs and use a supported authenticated CLI session. Never test overwrite behavior on your only copy of important data.

Archive boundaries

Dataset exports include configured entity data and asset information/files according to the command’s selection. A type-filtered export can contain a narrower asset set. Inspect the reference before assuming unreferenced files are included.

A dataset archive is not a full project backup of memberships, billing, credentials, integrations, and every AI resource. Preserve application configuration separately and document your recovery steps. Content deduplication does not mean repeated imports consume no requests or resources.