Actions15
Crawl URL
AI-generatedSummary
Crawls one URL in a browser through the Crawl4AI API, then returns structured page data such as Markdown, links, tables, and optional HTML or screenshot. Configure browser behavior and crawl settings as needed; output and filtering options control what content is extracted and included.
Inputs
- URL (required) — The page URL to crawl.
- Browser & Session — Optional browser configuration: profile headers, browser engine and channel, cookies, JavaScript, stealth mode, launch arguments, custom headers, headless mode, HTTPS handling, initialization scripts, resource-light modes, proxy, session and storage state, timeout, persistent context, user agent, and viewport. Use a session ID or persistent context when retaining browser state is needed.
- Crawl Settings — Optional crawl behavior: anti-bot handling, ad/CSS blocking, cache mode, robots.txt checks, CSS content selector, excluded tags, page JavaScript, session-only JavaScript mode, retries, link handling, wait conditions, and minimum word count. These settings affect page loading and content extraction.
- Output & Filtering — Optional response controls: choose raw or fit Markdown; include links, media, tables, or raw HTML; capture a screenshot or generate a PDF; fetch SSL certificate details; enable verbose response data; and configure content filtering or table extraction. BM25 filtering requires a user query; LLM filtering requires an instruction and LLM credentials. Pruning and table strategies have additional thresholds and chunking options.
Output shape
A single structured crawl result record for the URL.
The result includes crawl data such as Markdown and page metadata; optional fields depend on output settings and server data. Screenshot and PDF content, when requested, are returned as base64. The operation applies selected content-filter and table-extraction settings before formatting the result.