Ask your architecture
anything.
Your team explores the catalog in a browser. Your AI tools query it via MCP. Both get answers from the same catalog.
Built-in MCP server with 15 tools and 9 resources. Fully self-hosted.
Click a badge to see different queries
Your architecture, queryable from any AI tool
Every EventCatalog ships with a built-in MCP server. Connect Claude, Cursor, VS Code, or any MCP-compatible tool and query your architecture without leaving your editor.
Claude
MCP Server
EventCatalog
Query
What events does OrderService publish?
Response
OrderService publishes 4 events: OrderCreated, OrderUpdated, OrderShipped, OrderCancelled
15 tools, 9 resources, out of the box
Query domains, services, events, schemas, OpenAPI specs, AsyncAPI specs, channels, and more. No configuration needed.
Works where your engineers already work
Claude Desktop, Cursor, Windsurf, VS Code, or any MCP-compatible client. Connects in minutes.
One model, two audiences
Humans get a visual catalog. AI gets structured access. Both reason from the same model.
No additional infrastructure
Built into every EventCatalog instance. Nothing new to deploy or manage.
Stop wasting time. Start asking.
Your architecture knowledge shouldn't be locked in wikis, Slack threads, and people's heads.
Searching Confluence for 20 minutes
Ask your catalog and get the answer in seconds
Context-switching between editor and docs
Query your catalog directly from your IDE
Breaking changes discovered in production
Understand blast radius before you ship
Asking around to find who owns a service
Every resource has an owner. One query to find them.
Manually checking schema compatibility
See every consumer before you change a schema
New engineers lost in the architecture
New hires ask the catalog, not the team
Ask your catalog, not your colleagues
A built-in chat interface so your team can ask architecture questions directly in the browser. No Slack thread required.
Bring Your Own Models
Use OpenAI, Anthropic, or even browser-based models. You choose.
Answers grounded in your architecture
"What events does OrderService publish?" Answers come from your documented architecture, not hallucinated guesses.
Custom Prompts
Define your organization's best practices and standards as reusable prompts.
Make your architecture readable by any AI agent
EventCatalog auto-generates llms.txt and schemas.txt so AI agents can discover and understand your architecture without custom integrations.
llms.txt
Domains, services, events, and docs, structured so any AI agent can reason about your architecture.
schemas.txt
Every schema in one file: OpenAPI, AsyncAPI, JSON Schema, Avro, Protobuf. AI tools get full contract context.
Always current, zero maintenance
Regenerated on every build. When your catalog updates, your AI context updates too.
# EventCatalog Architecture ## Domains - Orders: Manages order lifecycle - Payments: Handles transactions - Inventory: Stock management ## Services - OrderService: /services/OrderService - PaymentService: /services/PaymentService - InventoryService: /services/InventoryService ## Events - OrderCreated: /events/OrderCreated - PaymentProcessed: /events/PaymentProcessed - StockUpdated: /events/StockUpdated
Self-hosted. Your data never leaves.
Your architecture docs, AI queries, and catalog data all stay on your infrastructure. No external calls. No vendor lock-in. Perfect for regulated industries.
Your infrastructure, your rules
Run on AWS, Azure, GCP, or your own servers. Static site, deploy anywhere you serve HTML.
No telemetry. No phone home.
The MCP server and AI assistant run entirely within your network. Nothing leaves unless you choose to send it.
Bring Your Own Models
OpenAI, Anthropic, local models, or browser-based AI. No vendor lock-in on the model layer either.
Make your architecture queryable today
Join the teams behind 31,000+ catalogs
Self hosting documentation and governance for your organization.
