遭LinkedIn起诉卖千万ARR业务，2周重启新工具半年月入1.5万美元

Lawsuits, injuries, and launches

I’m a software engineer and serial entrepreneur — 20+ years behind the keyboard. I've sold two companies. Proxycurl was by far the largest LinkedIn scraping API on the market, with $10M ARR. But then, LinkedIn sued Proxycurl in January 2025. We settled out of court, and I sold Proxycurl to a competitor. After the acquisition, I fractured the metatarsal bone in my right foot and found myself immobile — with Claude Code. Two weeks later, I prototyped and launched NinjaPear, a B2B company intelligence and data enrichment platform. That was six months ago. We did over $15k in gross revenue last month.

Picking the idea

I built NinjaPear because I knew the B2B enrichment data market so well. I wanted to build a Proxycurl with my own data instead of LinkedIn's. I wanted to build something that could scale for the next ten years without fear of the Microsoft legal war machine. So, I prototyped an API with LLM + Deep Web Research that provided rich and complete B2B data. The data is not sourced from LinkedIn, and it includes many data points that LinkedIn simply does not have, such as customer list data, competitor research, and product lists. I work on NinjaPear as a solo, full-time founder. All it took was two weeks of time and a Claude 20x Max subscription. But money isn't really an issue anyway. I’m an early investor in cryptocurrency, and I have sold two companies. My past successes allow me to work on NinjaPear indefinitely without needing to turn a profit — and like I said, it's already profitable.

The tech doesn't matter

In the post-AI world, the tech does not matter. My stack is boring: Python (type hints enable free bug catching via mypy, which you want because of AI), Typescript/Tailwind (frontend), Kubernetes, Redash for business intelligence, Grafana. I also own all the servers and co-locate them at my house, powering them with solar power.

Wrestling with churn

NinjaPear users pay for data with credits. Users purchase credits through subscriptions (which are cheaper) or pay-as-you-go top-ups. We rely on users integrating NinjaPear data into their workflow. We grow with them. The biggest challenge I'm experiencing right now is churn, and I have hypotheses about why it is happening. Our data processing is being compared to LinkedIn's, and it's slower because LLMs perform web-scale research in real-time. Product messaging is also an issue. I'm iterating constantly and working to fix both. There will always be a challenge. That’s what makes venture building fun! It’s the “IRL” game that I play. Every failure is a step closer to success.

A two-pronged GTM strategy

As far as marketing, I focus my attention on: Emailing my 20k+ newsletter subscribers, SEO efforts on the domain I retained despite selling Proxycurl, Posting on reddit — though this is more recent. I'll continue wearing the GTM hat for the next six months.

Getting okay with failure

These days, I take nothing personally — including failures and haters. In fact, I frequently tell myself, "Let’s set out to fail 10x." Then, I keep failing and recalibrating as I go. The cadence of the output is what matters. Just keep going. And keep failing. For example, a few days ago, I spent a couple of hours knocking out this article on Reddit, and I barely got any traction on it. Younger Steven would be very upset. Now, I know it means to try again and again, and eventually something will go viral.