We describe Infrahub as a data management platform rather than a source of truth (SoT) platform. That’s because data is destiny in automation, so the degree to which you can manage your data determines how far you can take your automation initiative. A "source of truth” in most cases is just a data repository. A data management platform focuses on giving engineers the structure and tools to make intent data capable of driving scalable, long-term automation impact for your organization.
At AutoCon 4, we're announcing our 1.0 MCP server that builds on Infrahub’s AI-ready graph data approach, and sharing new data management advances we have coming over the next few releases, including an overhaul of the Infrahub UI, and a new capability called Profiles, which are standardized patterns for managing object attributes and relationships at scale.
Infrahub’s MCP is GA, serving AI-ready data
The first highlight for AutoCon is the general availability of the Infrahub MCP server. MCP, which stands for Model Context Protocol, is an Anthropic-innovated protocol and framework that enables LLMs to securely and efficiently connect to and use external tools, data sources, and services such as Infrahub. That’s cool but let’s be honest. MCP is a beneficial interface, but it’s what sits behind that interface that makes it powerful or not.
Making data useful for LLMs is all about context. LLMs have limited contextual capacity to ingest data and make sense of it. Curating what you feed LLMs is critical to get great results.
When building an application around a database, the ability to provide context to LLMs is limited by the extent to which semantics are integral to the database's structure. If the database is just a container for arbitrary sets of data, and most or all of the semantics exist in the application code, then when trying to feed an LLM, the database itself won’t have much contextual meaning to serve. Without those semantics, LLMs will struggle to fulfill the intent of a request based on the data they receive. This is the reality for most applications, including traditional SoT tools.
One of the reasons we emphasized schema flexibility in Infrahub is to ensure that semantic context can be built into the data model. In many ways, Infrahub is conceived around the concept of a knowledge graph. Knowledge graphs are superior for feeding LLMs because, by definition, they map relationships between data objects, encompassing both infrastructure relationships and logical or business concepts. Infrahub’s flexible schema incorporates rich metadata, lineage, and temporal versioning relationships. This relationally rich approach offers LLMs the information they need to understand context and perform tasks with the greatest accuracy and precision.
Get an excellent introduction to the Infrahub MCP server in this blog post by Suresh Vina.
A new UI for contextual data interaction
In the coming weeks, Infrahub will be gaining a modern, relationship-aware user interface for infrastructure automation. Much like how CDPs unify fragmented customer data into a contextual, accessible, and actionable format, Infrahub reimagines how you interact with infrastructure data by eliminating rigid, form-based interfaces in favor of a responsive UI that layers data relationships, lineage, and schema context.
Whether you’re using a traditional SoT point product or a home-grown one your team has developed, it’s common for the data foundation to be a relational database. Home-grown SoTs project teams often struggle to find bandwidth to create much of a UI. For legacy commercial SoTs, the UI is typically designed as transactional forms wrapped around individual tables in the database. These legacy UIs are undoubtedly more usable than a CLI but are pretty limited.
We're working on an updated UI/UX paradigm, replacing form-heavy workflows that require multiple screens and saves with inline editing that allows faster updates.
But the redesign goes beyond guzinta guzouta. We’re also making metadata, lineage, and history far easier to see. This means you can explore and understand data structures, navigate objects, and easily see the underlying schema context for each object, item, and label.
This CDP-style experience is vital because it increases your team’s data confidence and speeds execution, so you can scale automation workflows, improve the reliability and accuracy of change management, and ultimately make a bigger positive impact on your business.
Our new UI improvements will be released in the very near future, so stay tuned!
Profiles: Standardized patterns for scalable object management
Modern infrastructure automation depends on having consistent, reusable models of how networks, systems, and services are built. Our next frontier in this effort is Profiles, a more powerful way to define and manage that consistency at scale. Profiles are not new to Infrahub, but over the next few months, you will see some powerful updates that take Profiles to the next level.
With Profiles, you'll be able to create reusable patterns that describe not only inherited attributes of objects but also their relationships. These reusable patterns let you apply (and reapply) standards across thousands of objects at once — based on scoping that makes sense in your environment — ensuring that every instance inherits the right values, relationships, and metadata automatically. It’s a foundational capability for maintaining data integrity and accelerating automation across large, complex environments.
If you’ve worked in network or infrastructure automation for any length of time, you'll undoubtedly have run across SoT support for object “templates.” In Infrahub, templates allow you to define consistent, standardized ways to create objects (such as devices and interfaces), including both fixed and immutable attributes, attribute variables, and even relationships.
That’s great, but how do you manage those variables declaratively after they’re created, at scale, in a schema-aware way? The answer until now was, you couldn’t. As a result, traditional SoTs and homegrown systems force teams to manually manage a wide range of object attributes and relationships, or rely on fragile scripts. When you’re managing thousands of objects, mistakes are made and drift occurs as you evolve infrastructure, designs, and services. At scale, such inaccuracies are very costly.
Our forthcoming updates to Profiles radically change the game. Profiles will become reusable, schema-driven patterns that declaratively encode attributes and relationships inheritance for arbitrarily scoped classes of objects. All relevant objects automatically inherit the attributes and relationships defined in the Profile, including when the Profile is changed.
Profiles eliminate duplication, enabling teams to create and manage thousands of objects from a single, metadata-rich blueprint — ensuring consistency, accuracy, and full traceability. Since Profiles are precomputed and stored rather than dynamically resolved, they will perform reliably at scale. And you can define multiple Profiles to be applied in a priority order to an object or object template, giving a lot of flexibility in how they’re used.
Here are a few ways Profiles will benefit your organization:
- Improved time-to-market due to faster infrastructure and network service delivery, supported by enhanced data accuracy and more consistent automation pipeline processes
- Improved team efficiency and higher team productivity due to less duplicated effort
- Better risk-adjusted loss reduction via more consistent application of security and compliance policies and superior auditability
- Reduced operational costs due to fewer error-driven outages and faster MTTR due to better data traceability
Run Infrahub at scale
A central theme you’ll see above and in many of our recent releases is making it easier for teams to run Infrahub at scale. For example, we recently released Infrahub Backup, a new tool that provides a unified, automated way to capture and restore your entire Infrahub using one consistent workflow. We also released the Infrahub VS Code extension, which brings real-time infrastructure-as-code productivity directly into your favorite IDE.
If you want to learn more or explore Infrahub, check out our GitHub, documentation, YouTube channel, hands-on labs, and join our Discord community. If you want to explore how you can get greater automation value from Infrahub, request a demo using the button at the top of the page.