🍮 crawlbrulee screenshots

you capture a screenshot as part of a scrape, through extract.screenshot. we don't inline the image — the response hands you a url you download separately, plus optional tiles for tall pages.

new to crawlbrulee? read the crawlbrulee skill first. for the surrounding request/response contract, read crawlbrulee-scrape.

two things to know up front

a screenshot always renders the page. you don't need require_js: true to get one — asking for a capture implies rendering.

in rare cases we can't capture one. what happens then depends on what else you asked for. if the screenshot was one output among several, it's best-effort: the screenshot field is left out of the payload and everything else still arrives — check the field is there before reading url off it. if the screenshot was the only output you requested, the call errors instead of returning an empty success: unsupported_screenshot_output (a 422) when the content type can't be screenshotted — json, plain text, markdown, xml — or a 500 when the capture itself failed. neither error is billed. the same rule applies to async jobs when you fetch the result.

options

extract.screenshot accepts: