It turns natural-language requests into Microlink calls: take screenshots, generate PDFs, extract metadata or readable text, detect video/audio sources, run Lighthouse audits, scrape custom fields with CSS selectors and more
