Scrape a JavaScript-heavy page with Firecrawl browser actions (wait/click/scroll/write) before capture

Documented steps

1. Auth: Authorization: Bearer fc-YOUR-API-KEY header against POST https://api.firecrawl.dev/v2/scrape (v2 API).
2. Build an `actions` array executed strictly in order before the final capture, e.g. [{"type":"wait","selector":"#app-loaded"},{"type":"click","selector":".accept-cookies"},{"type":"scroll","direction":"down"},{"type":"write","text":"search term"},{"type":"press","key":"Enter"},{"type":"wait","milliseconds":1500}].
3. POST /v2/scrape with body {"url":"...","actions":[...],"formats":["markdown"]} so the capture reflects the post-interaction DOM.
4. Optionally insert {"type":"screenshot"} or {"type":"scrape"} action steps mid-sequence to capture intermediate states; these land in data.actions.screenshots / data.actions.scrapes, separate from top-level data.markdown.
5. Raise `timeout` (default 60000 ms, max 300000) since action execution and waits count against the same request timeout.
6. Read data.markdown (or other requested formats) for post-interaction content, and data.actions.javascriptReturns if you used an executeJavascript action.
7. Other supported action types: click (optional `all` to click every match), press, pdf (renders a PDF of the interacted page).
8. Official docs: https://docs.firecrawl.dev/api-reference/endpoint/scrape ; https://docs.firecrawl.dev/features/scrape