The creator of NanoClaw, an open source AI agent platform with over 18,000 GitHub stars, says Google is ranking a fake website above his project’s real site.
In tests conducted on March 5, an impostor site ranked at the top of Google for the project’s own name. The real website, nanoclaw.dev, did not appear in the first several pages of results.
What’s Happening
Gavriel Cohen, a software engineer and former Wix developer, posted a thread on X describing the problem.
Cohen launched NanoClaw in early February as a security-focused alternative to OpenClaw, the viral open source AI agent platform. The project grew quickly. VentureBeat covered it, The Register profiled Cohen, and AI researcher Andrej Karpathy publicly praised the project’s architecture.
Around February 8, someone registered nanoclaw.net and created an auto-generated site scraped from the project’s GitHub README. Cohen said he didn’t have a website at the time because the GitHub repo was the project.
As the project gained press coverage, people kept contacting him about problems with “his” website. It wasn’t his.
He built the real site at nanoclaw.dev and then took several standard SEO and remediation steps. He linked it from the GitHub repo. He added structured data. He submitted to Google Search Console. He filed takedown notices with Google, Cloudflare, and the domain registrar. Publications covering the project linked to nanoclaw.dev.
As of March 5, the impostor site still ranked above the real one.
In his thread, Cohen wrote that the fake site is “showing factually wrong information about the project and falsifying its publication dates.” He called the situation “a live, active security risk” because the person running nanoclaw.net could replace the page content with malicious download links or a phishing page at any time.
The Hacker News thread about Cohen’s complaint reached 315 points and over 150 comments within hours.
Same Problem Across Search Engines
Hacker News commenters tested the same search on other engines and found the problem extends beyond Google.
One commenter reported that the fake site ranked #1 on DuckDuckGo and #3 on Kagi, while the real site didn’t appear on DuckDuckGo at all. Another found that Bing, Brave, Ecosia, and Qwant all showed the fake site in top positions. Mojeek was the only engine tested that ranked the real site and excluded the fake one.
Why This Matters
In the past, Google’s John Mueller said that copied content consistently ranking above the original may point to a site quality problem. Mueller suggested site owners reassess their overall quality if this keeps happening.
Cohen’s case tests that logic. His project has 18,000 GitHub stars, coverage from CNBC, VentureBeat, and The Register, a Karpathy endorsement, and a blog post that hit #1 on Hacker News. Every social profile and the GitHub repo itself point to nanoclaw.dev. On its face, many of the visible signals appear to favor the real site.
The fact that Hacker News commenters reported similar results across multiple search engines suggests something deeper than a Google-specific bug. One possible factor is timing, as the fake site appears to have been indexed before the real site launched.
For anyone building a new product, the key takeaway here is to reconsider the right time to register a domain. Cohen focused on shipping code before building a website. That’s standard open source practice, but search engines indexed the impostor first, and correcting that after the fact proved harder than any of the recommended steps suggest it should be.
Looking Ahead
Cohen has not indicated whether Google responded to his takedown requests. One SEO practitioner in the Hacker News thread offered concrete advice, including mapping the fake site’s backlinks and contacting publications that accidentally linked to the wrong domain.
The situation remains unresolved. Google had not commented at the time of publishing.
Featured Image: Elnur/Shutterstock
