firecrawl-scrape

Run firecrawl scrape with the quoted URL and an output flag, for example firecrawl scrape "<url>" -o .firecrawl/page.md. Add --only-main-content to strip navigation and footers from the result.

Use the --wait-for option with a millisecond delay, such as --wait-for 3000, so the page's JavaScript finishes rendering before extraction. If the page requires clicks or form interaction, escalate to the firecrawl interact skill instead.

Yes, passing multiple URLs to firecrawl scrape processes them concurrently and saves each result as markdown under the .firecrawl/ directory. In this mode the -o flag is ignored and only markdown output is kept.

Prefer firecrawl scrape whenever the user provides a URL and wants its content, since it returns clean LLM-optimized markdown and handles JS-rendered pages. Use the --query option only for a single targeted answer, as it costs extra credits.

Requesting multiple formats such as --format markdown,links produces JSON output rather than raw markdown. In multi-URL mode, if markdown was not requested, the entire JSON response is written into the .md file.