How can StackBlitz WebContainers efficiently paginate large datasets fetched from external APIs without exceeding browser memory limits?

Goal: allow a Node.js app running in StackBlitz to retrieve and display paginated results from a large external dataset while staying within browser memory constraints.

Constraints: WebContainer runs in a single tab, limited heap, SharedArrayBuffer requirement, network proxied requests may add latency; need to bound query size and avoid loading entire dataset into memory.

Uncertainty: whether to implement pagination at the API level (limit/offset or cursor) or client-side buffering, and how to handle state persistence across terminal reloads.

Which pagination strategy (limit/offset vs cursor) minimizes memory usage in a WebContainer environment?

How should the client-side state be managed to survive hot reloads without refetching?

What maximum page size is safe given typical browser tab memory limits for StackBlitz?