NetBox is the semantic map that provides critical context for network and infrastructure AI, a realm of rapidly increasing investment and engagement from AIops companies, global IT leaders, and infra engineers in every industry. At NetBox Labs, we’ve been working hard to unlock the value of NetBox’s semantic context to support AI and agentic workflows.
Since we released the NetBox MCP server in March, the response from the networking and infrastructure community has been remarkable. What started as an open-source tool to connect AI agents with NetBox’s infrastructure data has evolved into something much larger: the foundation of an emerging ecosystem for infrastructure AI. At the same time, we’ve built more high-value agentic capabilities around NetBox, like the experimental NetBox Operator AI operations platform, and “here and now” products like NetBox Copilot, which is nearing GA as a result of massively successful private and public preview programs.
The NetBox MCP server numbers tell part of the story. We’re seeing hundreds of downloads monthly, with dozens of teams experimenting to address their specific use cases and demonstrating paths to valuable AI outcomes with NetBox. Enterprise architects are incorporating these patterns into their broader AI strategies, recognizing that NetBox’s semantic map provides something critical that’s been missing from most infrastructure AI efforts: context.
The real indicator of impact is what people are building with NetBox MCP. Across an array of community blog posts, like this demonstration by Microsoft MVP Simon Painter, we’re seeing a common thread we have seen with NetBox itself. Teams are using NetBox MCP to connect with LLMs and agents to open access to NetBox data for non-technical users. Security teams, compliance groups, capacity management teams, and other stakeholders can now ask natural language questions about infrastructure and quickly get answers. This democratization of infrastructure data is reducing operational overhead as engineers shift focus from answering routine queries to higher-value strategic work.
What’s also encouraging is seeing production-ready deployments that demonstrate genuine value at enterprise scale. Network engineers are successfully deploying distributed MCP server environments with sophisticated orchestration capabilities. In one example, an engineer demonstrated using MCP with NetBox to populate 68 interfaces in under a minute, showcasing the kind of acceleration that makes this approach viable for meaningfully scaled operations.
These aren’t proof-of-concept demos. These are production-capable systems handling real infrastructure workflows, guided by NetBox’s semantic understanding of the environment.
Perhaps the strongest validation comes from engineers at leading networking vendors who are actively incorporating the NetBox MCP server into their AI experimentation workflows. Cisco engineers have publicly documented their exploration of the tool as a foundational component for AI-augmented networking.
One Cisco engineer captured why this matters: NetBox’s semantic map enables AI agents to “have access to the same information as human network engineers” for powerful cross-platform operations. When you think about it, that’s exactly what’s been missing from most infrastructure AI efforts. Agents need to understand not just individual data points, but how everything connects and relates.
Another noted that “having access to source-of-truth data will be key to any Agentic NetAI work we undertake.” This recognition from industry leaders validates what we’ve believed from the beginning: effective infrastructure AI requires a semantic foundation, not just more automation.
The open, composable nature of the MCP protocol is proving crucial to adoption. Teams can combine NetBox’s semantic map with other agentic tools for accessing network infrastructure, gathering telemetry, and orchestrating operations. This composability enables powerful cross-platform workflows for network and infrastructure operations, all guided by the strategic context that lives in NetBox.
We’re seeing teams build integrated workflows that span multiple systems and vendors, unified by the semantic understanding that NetBox provides. The NetBox MCP server doesn’t lock teams into a single approach or platform. Instead, it opens up possibilities for integration across the entire infrastructure stack.
The NetBox MCP server is not a standalone solution. It’s part of a broader NetBox AI ecosystem, which now includes NetBox Copilot as a fully interactive, embedded agent powered by our Nitro AI platform.
Where the NetBox MCP server provides the foundation for custom AI development, NetBox Copilot delivers immediate value with foundational agent capabilities that will be freely available to the entire networking and infrastructure community. NetBox Copilot is lightweight and easy to implement, enabling teams to get started with agentic AI workflows right away, leveraging NetBox’s semantic map to accelerate day-to-day operations and unlock the value of NetBox data for teams beyond engineering. We’re opening up more spots from the NetBox Copilot public preview waitlist daily, incorporating enterprise-grade features for data governance, and heading toward GA later this year.
NetBox is deployed across nearly every sector and scale of enterprise, from Fortune 500 companies and government agencies to industrial infrastructure and AI datacenters. This widespread deployment means the NetBox MCP server and the broader NetBox AI ecosystem can be adopted incrementally, unlocking infrastructure AI benefits without requiring teams to replatform existing systems.
The semantic map approach solves a fundamental problem in infrastructure AI. Without structured understanding of how things connect, where dependencies exist, and what “correct” looks like, AI agents make costly mistakes. But when grounded in NetBox’s comprehensive data model, these same agents become remarkably effective at complex infrastructure reasoning and operations.
What we’re seeing with the NetBox MCP server represents more than just a successful open-source project. It’s evidence that the industry is ready for a new control plane for AI-augmented networking. The combination of NetBox’s semantic foundation, the MCP protocol’s composability, and the community’s innovative implementations is creating something entirely new in infrastructure management.
The barrier to building effective infrastructure AI has been dramatically lowered. Teams no longer need to start from scratch or build complex integrations to give AI agents meaningful context about their environments. The NetBox MCP server provides that foundation, and the community is building remarkable things on top of it.
This is just the beginning. With initial learnings from the community, NetBox Operator, and NetBox Copilot, we’ll be refining the NetBox MCP server to be even more effective for agentic AI use cases. As more teams adopt semantic approaches to infrastructure AI, we expect to see even more sophisticated use cases emerge. The foundation is solid, the community is engaged, and the possibilities are expanding rapidly.
If you want to get hands-on experience building AI agents for infrastructure management, I’m hosting a workshop where we’ll walk through creating your own AI agent using the NetBox MCP server and modern AI tools. You’ll learn how to connect AI agents to NetBox’s semantic map, build practical workflows, and understand the patterns that are driving successful infrastructure AI implementations.
We’ll cover everything from setting up the NetBox MCP server to designing agent workflows that leverage NetBox’s rich data model. Whether you’re just getting started with infrastructure AI or looking to enhance existing automation, this workshop will give you practical skills you can apply immediately.
Register for the workshop here and join the growing community of teams building the future of infrastructure AI.