Firecrawl raises $75M Series B to build a knowledge library for AI agents
What's the deal? Firecrawl, a US startup founded by Brazilian entrepreneur Nicolas Silberstein Camara alongside Caleb Peffer and Eric Ciarla, has raised $75 million (about US$102.3M) in a Series B round led by Smash CapitalDealroom has a profile for this one. Try Dealroom →. Altos Ventures, Nexus Venture Partners, Y Combinator, Freestyle, and Offline VenturesDealroom has a profile for this one. Try Dealroom → also took part.
What's the endgame? Founded in 2024, Firecrawl builds open-source tools to extract, structure, and process data from the internet for AI developers and agents. The money will fund its expansion and the development of Alexandria, a platform to gather knowledge from many sources for AI models and agents.
Alexandria aims to pull together scientific research, code, public documents, and real-time data. It already draws on public papers in medicine, biology, and AI; knowledge from 70 million code bases; and tens of millions of government documents, including court records and filings from the US Securities and Exchange Commission.
Why now? The round lands a year after Firecrawl's US$20.7M Series A, led by Nexus and Y Combinator. The company now serves more than 1.5 million developers and 150,000 companies worldwide, including Shopify, Apple, Lovable, and Canva.
Growth in Brazil has been steep. In the first three months of 2026, Firecrawl grew 141% in the country, and within five months it had more than double its paying customer base from all of 2025 — expanding at roughly 290% a year.
"Receiving a new investment round just one year after the Series A reinforces investors' confidence in the infrastructure we built and in Firecrawl's ability to meet the growing demand for AI data," said Silberstein Camara, chief technology officer and co-founder (translated from Portuguese).
The signal: The raise sits in the top 5% of all-time Series B rounds for US enterprise software startups, a marker of how much capital is flowing toward the plumbing that feeds AI models. As models grow more capable, the bottleneck shifts to the scattered data they need — and Firecrawl is betting on becoming its central library.
Read more: exame.com
Image credit: Firecrawl