AI/LLM Context Foundation

The problem​

AI tools are transforming how we build software, but they are missing something critical. They need your business context. Your domains, your ubiquitous language, your service boundaries, your systems, your organizational patterns and rules. Without this context, AI assistants generate generic advice that does not align with your actual architecture.

Developers spend their time repeatedly explaining their architecture in every conversation. AI suggests patterns that contradict your domain boundaries. It recommends building features that already exist under different names. It generates code that does not understand which services own what data or functionality. The AI has the capability to help, but it lacks the foundational knowledge of your specific organization.

Where is the AI going to get that information? Your architecture lives in scattered documentation, tribal knowledge, and code comments. The AI cannot access it, cannot understand it, and cannot reason about it. You need a single source of truth that both humans and machines can consume.

The solution​

EventCatalog provides foundational context for AI and LLMs. Every EventCatalog instance comes with a built-in Model Context Protocol (MCP) server that exposes your architecture documentation to any AI assistant. Connect your preferred AI tools, bring your own models, and give them direct access to your living documentation.

The MCP server transforms your Markdown-based documentation into a queryable knowledge base. Ask EventCatalog questions directly from your IDE or any MCP-compatible client. What events do we have? What services do we have? Who owns them? What domains exist in our organization? The AI gets accurate answers based on your actual architecture, not generic assumptions.

Because EventCatalog is powered by Markdown, the quality of your documentation directly determines the quality of AI responses. Document your architecture for humans, and machines automatically benefit. The technical implementations, producers and consumers, domain boundaries, and ownership structures you already maintain become the context that makes AI truly useful.

Beyond text: Diagrams as queryable knowledge​

EventCatalog takes this further with its visualization capabilities. The automated node graphs and visualizations that help humans understand relationships get transformed into Mermaid diagrams that LLMs can read and reason about. When you ask a question about a visualization, the AI understands the relationships between services, events, and domains because the diagram itself becomes part of the context.

This means AI can help you understand not just what exists, but how components relate to each other. It can trace event flows, identify potential impacts of changes, and recommend architectural patterns that align with your existing system design.

Real workflows enabled​

The market is changing. Organizations are adopting AI tools at an accelerating pace. EventCatalog positions you to take advantage of this shift by providing the architectural context that makes AI assistants genuinely valuable.

For onboarding, new team members ask questions and get accurate answers immediately. What things do we have in our architecture? What produces this event? I am looking for this specific event to integrate with. The AI becomes an always-available guide to your system, drawing directly from your maintained documentation.

For reducing redundant work, developers planning new features can ask the AI to find similar existing events or services. Instead of accidentally rebuilding functionality that already exists under a different name, they discover what is already available. The AI can make better architectural recommendations because it knows what systems you already have and how they are structured.

The AI understands ownership and team boundaries because that information lives in EventCatalog. It can reference your actual service names, event types, and domain language. It speaks your organization's ubiquitous language because that language is documented and accessible.

How this can help you​

Documentation for humans and machines. That is the vision. EventCatalog becomes the heart of how your organization shares knowledge, not just among people, but with the AI tools you increasingly rely on.

You get AI assistants that understand your specific context, not just general software patterns. They help you work within your architectural constraints and standards rather than against them. They surface relevant existing components instead of suggesting you build everything from scratch.

Your documentation investment pays double returns. The effort you put into maintaining EventCatalog helps your human team members and simultaneously powers your AI tools. The Markdown format that makes documentation maintainable also makes it machine-readable. The diagrams that help visualize complexity become queryable knowledge graphs.

As AI adoption grows across your organization, EventCatalog ensures those tools have the foundational context they need to be genuinely useful. Connect any MCP client, bring your own models, and give them access to the architectural knowledge that makes the difference between generic advice and truly helpful assistance.