ShotGrid Custom Entity Pagination Across API and UI Compatibility Boundaries

Custom entity views in ShotGrid must render consistently in both the web interface and external integrations. The web UI appears to handle large result sets through automatic lazy loading, while the Python and REST APIs require explicit pagination logic using filters, sorts, and limits. A decision is needed on whether to standardize on offset-based paging or investigate cursor-based continuation tokens for custom entities, particularly before datasets grow beyond a few thousand records.

The uncertainty centers on documented behavior at this compatibility boundary. Large offset values risk server-side performance degradation, and API rate limiting may interrupt high-volume page requests. Additionally, changes to custom entity schemas can invalidate hard-coded field filters inside pagination scripts, creating a maintenance risk that differs from the UI's managed retrieval path.

Before committing to a pagination strategy, the supported mechanics and safe thresholds need confirmation.

Does the ShotGrid REST API return a continuation token or offset marker for custom entity find requests, or is manual offset arithmetic the only documented approach? When a custom entity schema evolves, what is the supported method for updating paginated field filters without breaking existing API consumers? At what record volume does ShotGrid documentation recommend moving from basic offset paging to a filtered or sorted paging pattern to avoid performance issues?