Skip to main content
Set full_page: true to capture the scrollable page rather than only the initial viewport.
Full-page images cost 2 credits on successful completion. PDF output costs 3 credits whether or not full_page is enabled.

Select an algorithm

by_sections is the default. The engine scrolls downward, captures visible sections, and assembles the output. This lets it capture content that appears during scrolling. native uses the engine’s native full-page capture path after scroll preparation. Test both algorithms against your target when sticky navigation or scroll-dependent effects affect the result.

Control scrolling

More waiting can help a site load lazy content, but also increases work. Start with defaults and adjust only when the output shows missing content. Use a readiness selector when the application has a dependable loaded state.

Detect truncated output

Infinite-scroll sites do not have a stable end. The engine uses deployment-configured height and scroll limits so a capture can terminate. Inspect render_diagnostics.truncated and stopped_reason when they are returned. A successful artifact can still be shorter than the target’s eventual infinite document. For a fixed report, use min_capture_height or required text checks to enforce expectations. For a feed, define a finite section to capture with selector instead of assuming that all future items can be included. Full-page capture can change page state as it scrolls. If exact animation frames matter, validate the output on your page with reduce_motion configured appropriately.