en.Wedoany.com Reported - Israeli AI agent security startup NanoCo has completed an oversubscribed $12 million seed funding round for its security-focused AI agent platform, NanoClaw. The round was led by Valley Capital Partners, with participation from several prominent angel investors and strategic backers, including companies such as Docker, Vercel, Monday.com, and Slow Ventures, as well as individual investors like Hugging Face CEO Clem Delangue, Auth0 co-founder Matias Woloski, and former Airbnb CTO Vanja Josifovski. Among them, Clem Delangue proactively reached out to the founders via social media and ultimately decided to invest. In just a few weeks, NanoClaw garnered approximately 22,000 stars on GitHub, with developer community attention on the tool rapidly surging.
NanoCo co-founder Gavriel Cohen revealed that after gaining attention and endorsements from prominent AI figures including Andrej Karpathy and Singapore's Foreign Minister, the company rejected acquisition offers from multiple investors, including one valued at up to $20 million. After consulting with other entrepreneurs, Gavriel Cohen received key advice: the value of an open-source project grows exponentially as its community expands, and giving up company control would be premature. Choosing to develop independently, the funds will be used to expand the engineering team and develop enterprise-grade features. The company has already begun signing enterprise clients and providing implementation support services.
NanoClaw was founded by the Cohen brothers in early 2026 as a secure alternative to the widely popular OpenClaw framework. Gavriel Cohen discovered that OpenClaw had downloaded and stored his WhatsApp messages in unencrypted plain text on his computer, a finding that prompted him to build a more secure alternative within weeks. The resulting NanoClaw is a minimalist and auditable framework with a codebase of approximately 500 lines, compared to OpenClaw's estimated 800,000 lines of code. Its core technological innovation lies in its security architecture—running AI agents inside micro-VMs via Docker sandboxes, preventing a compromised agent from accessing the host or other agents' environments.
Despite its security architecture, NanoClaw is not invulnerable. Attacks such as prompt injection can still guide AI-generated code to include backdoors or malicious dependencies. The platform's rapid rise, along with developer demand for auditable, transparent AI agent frameworks, is driving security to gradually become a new industry standard, with other AI companies also beginning to offer security solutions like private deployment. NanoClaw plans to keep its core tool free, generating revenue by selling services and providing embedded engineering support, with technical staff working closely with clients to implement the product.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
