Import and Export

Sink's authenticated JSON import and export APIs move link records between compatible instances. They are separate from KV-to-D1 migration and R2 snapshots.

Export

Export reads authoritative D1 links, including expired records, in cursor-paginated JSON pages. Continue with the returned cursor until list_complete is true. The page size is controlled by the batch setting in configuration.

The portable record includes link identity, timestamps, expiration, routing, tags, OpenGraph fields, safety and cloaking flags, and a protected password representation. It does not expose plaintext passwords.

Import

Import accepts at most half the configured export page size per request. The complete request is schema-validated before item processing, then each item reports success, skip, or failure.

Important behavior:

• Expired records are allowed and retained.
• Active slug conflicts are skipped rather than overwritten.
• Slugs are normalized according to the destination instance's custom-slug case setting.
• Portable password hashes from Sink exports can be imported without knowing the plaintext password.
• Dashboard-masked password placeholders are not valid plaintext passwords.
• Successful writes go to authoritative D1 and then update KV as a best-effort cache.

Import is not a full database restore: it does not recreate schema, tombstones, migration runs, or the KV migration marker.