OptionalcacheCache TTL in milliseconds (defaults to 1 hour)
OptionalskipWhether to skip cache and force fresh previews (defaults to false)
OptionalcacheCache backend for preview results.
The default backend is process-local memory. It is safe for tests, local development, and single-process CLIs, but it is not shared across pods and does not survive restarts. Production services that cache preview URLs should provide a shared backend such as Redis or Postgres, and the cached URL itself must still resolve to durable storage rather than process-local assets.
Options for batch preview generation