Firecrawl raises $75M to turn web crawling into an agent data library

Why it matters

Firecrawl is moving from selling crawling infrastructure to controlling the catalogue agents query, while promising data owners a share of the resulting economics.

Firecrawl, founded by Caleb Peffer, Nicolas Silberstein Camara and Eric Ciarla, announced a $75 million Series B led by Smash Capital on September 22nd and launched Alexandria, a service that gives AI agents access to specialized datasets through one interface.

RuntimeWire scooped the financing on September 14th by monitoring SEC filings. A Form D from Firecrawl's legal issuer, SideGuide Technologies Inc., showed $82,063,463 of preferred stock sold to seven investors, with the first sale recorded on August 31st. The filing did not identify the investors, disclose a valuation or assign a round label. Firecrawl's announced Series B is about $7.1 million below the gross amount reported to the SEC, and neither disclosure explains the difference.

https://x.com/firecrawl/status/2102426246235775068

The financing moves the San Francisco company beyond its original business of turning websites into structured, model-ready data. Alexandria combines more than 100 data providers, site-specific connectors and Firecrawl's own indexes, according to Firecrawl's announcement on X. The initial catalogue covers research papers, code documentation, US property listings, professional profiles, products and prices.

Peffer, Camara and Ciarla arrived at that product from a problem inside their previous companies. The three founders built SideGuide, a developer-education product, and Mendable, an AI search service that needed to collect and clean web data. Firecrawl emerged from the ingestion infrastructure they built for Mendable. Peffer and Ciarla graduated from the University of New Hampshire in 2022, while Camara earned a computer science degree there in 2021. Their work together dates to their student startup years at UNH.

From crawler API to data catalogue

Firecrawl's crawler handles the messy work between a public website and an AI application: fetching pages, rendering JavaScript, extracting content and returning structured output. Alexandria pushes Firecrawl further into the data supply chain. Developers can ask one service for records from purpose-built datasets rather than sending an agent across the open web and accepting whichever pages a search tool retrieves.

Firecrawl uses housing and employment searches to illustrate the difference. An agent could compare a large set of job openings paying more than $150,000 with apartments renting for less than $4,000 a month, rather than work from a small group of search results. Alexandria is available through Firecrawl's command-line tool, Model Context Protocol integration and API.

Firecrawl says Alexandria produced answers that scored 21% higher than built-in web tools across more than 10 categories. Firecrawl ran roughly 1,000 catalogue-based questions using the same models and prompts, then used blind AI judging to grade the answers. That result comes from an internal evaluation, rather than an independently administered benchmark, and measures performance on questions designed around catalogue data.

The financing was led by Smash Capital, with Altos, Nexus Venture Partners, Y Combinator, Freestyle and Offline Ventures participating, according to reporting by Exame. The round's valuation was not disclosed.

Firecrawl previously raised a $14.5 million Series A led by Nexus in August 2025. That round also included Y Combinator, Shopify CEO Tobias Lutke, Postman CEO Abhinav Asthana, Mux founder Matt McClure and Zapier. Firecrawl went through Y Combinator's Summer 2022 batch.

Paying for the inputs

Peffer and his co-founders are also using Alexandria to address a conflict embedded in the agent economy. AI products need current, structured information, while the publishers, researchers and data owners producing that material increasingly want payment and control over automated access.

Firecrawl says part of the Series B will fund payments to researchers, publishers, experts, creators and other data providers whose material is distributed through Alexandria. The provider program currently takes the form of a waitlist, while Alexandria is already available to developers. Firecrawl's ability to recruit high-value sources, establish useful licensing terms and normalize incompatible datasets will determine whether Alexandria becomes infrastructure or another wrapper around existing providers.

Firecrawl claims 1.25 million developers, 150,000 companies, 5 billion requests served and 2.5 million weekly downloads across npm and PyPI. Firecrawl also reports 150,000 stars for its open-source software on GitHub. Those figures describe distribution and usage, rather than revenue or paid adoption.

Alexandria puts Firecrawl into a crowded group of companies supplying search, crawling, browser automation and data infrastructure for agents, including Exa, Tavily, Parallel, Browserbase, Apify, Bright Data, Oxylabs and Crawl4AI. Firecrawl's wager is that its crawler can serve as the entry point to a broader catalogue business. The $75 million round gives Peffer, Camara and Ciarla the capital to secure data providers before AI developers settle on a default access layer.