Over the last several years, NetBox has established itself as a widely used open-source tool for understanding and modeling networks.
NetBox Labs, the lead commercial sponsor behind the open-source project, has continued to push ahead and expand beyond just the core NetBox project. The company was spun out of DNS platform provider NS1 in 2023. NS1 was subsequently acquired by IBM. NetBox Labs provides commercially supported services for NetBox including cloud and enterprise offerings.
Back in Nov. 2023, the company announced previews of its network drift configuration technology known as NetBox Assurance. That technology became generally available on April 2. NetBox Labs is expanding even further with a series of generative AI capabilities.
NetBox is now testing out Model Context Protocol (MCP) server implementation, llms.txt support, and AI-powered operational APIs – all aimed at transforming how network teams interact with infrastructure data.
“We’re all in on AI,” Kristopher Beevers, CEO of NetBox Labs, told Network World. “The surety of network management and infrastructure management is going to change substantially as a result, especially of agentic AI this year.”
NetBox: From documentation to operational intelligence
At its core, NetBox functions as the authoritative system of record for network infrastructure configuration, serving organizations ranging from small teams to Fortune 100 enterprises. The platform documents IP addressing, network topology, device configurations, rack layouts and other critical infrastructure elements that network teams need to maintain visibility across their environments.
“NetBox is intent,” Beevers explained. “This is where network teams are documenting ‘Here is what the network and the infrastructure should look like.’ Think of intent as what is in NetBox.”
With the general availability of NetBox Assurance announced this week, the platform now extends beyond documentation to address the persistent challenge of configuration drift–when live network states deviate from documented intentions.
“The problem we’re tackling is operational drift,” Beevers elaborated. “Drift is when the actual network deviates from your intent. NetBox is intent, discovery is observation, and assurance is reconciling these things.”
The Assurance solution leverages NetBox Lab’s agent-based discovery architecture, which differentiates it from traditional monolithic network discovery tools.
“We decided, from the outset, to build our discovery in an agentic approach… these little, tiny agents that you can put wherever you want in the network, and then orchestrate them as a fleet,” Beevers detailed.
This architectural approach has proven particularly valuable for organizations with segmented networks. He added that there are all kinds of networks that have never had any ability to be discovered by typical tools in the past, because they’re heavily segmented or in isolated pockets for security and compliance reasons.
These developments arrive as NetBox has become essential infrastructure for AI development environments themselves.
“Every AI infrastructure is built around NetBox,” Beevers noted. “Everybody in this infrastructure space is using NetBox to address the demand for GPU infrastructure and AI data center infrastructure.”
AI infrastructure and tooling
NetBox Labs is positioning itself at the intersection of network management and AI with several new capabilities, starting with a Model Context Protocol (MCP) server implementation. This allows large language models to directly query and interact with NetBox’s infrastructure data.
“Exposing the data in your systems of record, like NetBox, to LLMs through tool use is a huge unlock,” Beevers emphasized. “Coupled with everything you’re hearing and seeing on agentic AI and long-running processes, our vision is transforming how network operators function in the NOC.”
The company also added llms.txt support, ensuring AI assistants have up-to-date knowledge about NetBox capabilities beyond their training cutoff dates, addressing the knowledge obsolescence issue that affects all LLMs.
Two AI-powered operational features demonstrate NetBox’s practical application of these technologies: an Enrichment API that contextualizes alerts and tickets with relevant infrastructure data, and an upgrade risk analysis tool that identifies potential integration issues during NetBox upgrades.
“Our enrichment API lets you send raw data like tickets or alerts to the API, and we use AI to find relevant context from NetBox so you can accelerate troubleshooting and automate remediations,” Beevers explained.
Agentic AI vision for network operations
Looking forward, NetBox Labs envisions autonomous AI agents transforming network operations centers, combining observability data with NetBox’s infrastructure knowledge.
“If you think about what a network operator does in the NOC all day, we really think that agentic LLMs coupled with tool use, with respect to NetBox and other operational tools like metrics and logging, are going to be capable of significantly improving operations this year,” Beevers predicted.
This vision may fundamentally alter network management workflows.
“If you are not experimenting with and investing in figuring out how this is going to impact your enterprise network operations right now, you’re going to be behind in a year,” Beevers warned.