Firecrawl builds web-data tools for developers creating AI agents, handling the search, page loading and extraction work between an agent’s request and the information it needs. SiliconANGLE reports that the company announced a $75 million Series B led by Smash Ventures. The financing accompanies Alexandria, a new cloud service that combines web content with outside and company-curated datasets.
A browser is designed around a person who can wait for a page to load, scroll down or click through a form. Automated collection is less forgiving: content that appears in stages can be missed when a scraper downloads a page too early. Firecrawl’s platform is meant to manage those interactions before passing the resulting data to an application.
Developers can use Firecrawl’s search engine to filter pages by criteria such as creation date or file type. The company says its smart-wait feature can hold collection until dynamically loaded content appears, while prompt-based controls let the software scroll, click interface elements or submit forms without a developer scripting every action.
The platform also extracts information from PDFs, Word documents and other files, and it can watch pages for changes. One use described by SiliconANGLE is monitoring a rival software provider’s pricing page, illustrating how the same collection system can support continuing business research rather than a one-time search.
Alexandria extends that workflow beyond the open web. Firecrawl is combining web data with third-party sources and curated collections, including scientific-paper abstracts and technical code and documentation. An AI coding assistant, for example, could pair a forum example with an explanation from a technical dataset. Firecrawl says a single application interface reduces the need to build a separate connector for every source.
According to SiliconANGLE, much of the new capital will support the addition of more third-party information to Alexandria. Firecrawl also plans to introduce a self-service content-licensing system, addressing the supplier side of a service whose usefulness depends on how much relevant data developers can reach through the same interface.