NanoCo Raises $12M Seed Round to Bring Secure AI Assistants Into the Enterprise

Enterprise AI is rapidly evolving beyond chatbots and copilots into systems that can take action inside real business workflows. NanoCo is positioning itself at the center of that shift with a $12 million seed round aimed at scaling its enterprise AI assistant platform built on top of its open-source framework, NanoClaw.

The funding round was led by Valley Capital Partners and included participation from Docker, Vercel, monday.com, Slow Ventures, Clutch Capital, and Factorial Capital, alongside angel investors including Clem Delangue.

According to the company, NanoClaw has already surpassed 250,000 downloads and accumulated nearly 29,000 GitHub stars since launching earlier this year.

From AI Chatbots to AI Employees

The broader AI industry is increasingly focused on agents that can execute tasks rather than simply answer questions. NanoCo’s platform is designed around that idea.

Instead of operating as a standalone assistant in a browser tab, NanoCo embeds AI agents directly into workplace tools like Slack and Microsoft Teams. The assistants are designed to interact with internal systems, customer data, developer tools, and company knowledge bases while adapting to individual employees and workflows over time.

The company says its assistants can help legal teams draft contracts, developers review code, and sales teams manage customer accounts, while maintaining persistent memory across conversations and projects.

That persistent context is becoming increasingly important as enterprises experiment with long-running AI workflows tied to proprietary internal knowledge. The industry is moving toward increasingly complex AI systems capable of interacting across internal tools such as Slack, GitHub, Gmail, Jira, and Confluence.

Security Has Become the Core Battleground for AI Agents

NanoCo’s rise comes as enterprises grow more concerned about the risks associated with autonomous AI agents.

Many current AI agent frameworks prioritize capability and flexibility, but enterprises are increasingly demanding auditability, isolation, and governance before allowing AI systems access to sensitive internal infrastructure.

NanoClaw attempts to address those concerns through a security-first architecture centered around container isolation. Every AI agent runs inside its own sandboxed environment powered by Docker infrastructure, while credentials are injected at runtime through secure gateways instead of being directly exposed to the agent itself.

The system also supports policy enforcement, approval workflows, and full audit trails for enterprise compliance teams. Sensitive actions can require human approval before execution, while organizations retain control over permissions and access policies.

That emphasis on isolation reflects a growing trend across enterprise AI infrastructure. Security teams are increasingly treating AI agents less like chat interfaces and more like privileged software processes that require containment and governance similar to cloud workloads.

Open Source Momentum Is Accelerating Enterprise Adoption

NanoClaw’s rapid adoption also highlights the growing influence of open-source AI infrastructure in enterprise environments.

The framework positions itself as a lightweight alternative to larger agent systems by focusing on a smaller, more auditable codebase with fewer dependencies.

That simplicity appears to be resonating with developers and infrastructure teams increasingly wary of opaque AI systems that are difficult to audit or secure. The project’s partnerships with Docker and Vercel further suggest that infrastructure providers see AI agents as an emerging workload category that will require entirely new deployment and governance layers.

The larger shift may ultimately extend beyond AI assistants themselves. As enterprises begin deploying persistent AI agents across departments, the infrastructure required to safely manage permissions, memory, workflows, approvals, and observability could become one of the defining software categories of the next decade.

For companies like NanoCo, the opportunity may depend less on building the most advanced model and more on building the systems enterprises are actually willing to trust.