EventCatalog Slack integration
Your teams can now ask EventCatalog about your architecture, events, services, and schemas directly in Slack.
You know the pattern. Someone on your team needs to know which services consume the OrderCreated event. They check the wiki. Not there. They search Confluence. Wrong version. They ping three different people in Slack asking if anyone knows.
We've made this much simpler. @mention the EventCatalog bot in any Slack channel or post in a dedicated architecture question channel. The bot reads your EventCatalog documentation and provides accurate, grounded answers in seconds.
Example of the EventCatalog Slack Bot in action asking about the OrderCreated eventFinding architecture information is hard
Your team spends hours every week hunting for architecture information. The knowledge exists somewhere. It's just scattered across wikis, Confluence pages, outdated diagrams, and people's heads.
A developer needs to understand an event schema to fix a bug. They check the docs. Maybe it's there, maybe it's not. Maybe it's three versions old. So they go to the schema registry, the schema is there, but it has no business context, no ownership, no history, no visualizations. So they ask someone who worked on it last month.
They waste 30 minutes tracking down the current schema from someone who worked on it last month... the list goes on.
This happens constantly. Every new team member. Every cross-team collaboration. Every time someone touches code they didn't write.
What's new
The EventCatalog Slack Bot connects directly to your EventCatalog instance through the EventCatalog MCP server. Your team asks questions in Slack. The bot queries your catalog and responds with answers grounded in your actual documentation.
The bot is self-hosted. You own your data and control your deployment. It uses Socket Mode, so no public URLs or webhooks needed. Configure it with your AI provider of choice (Anthropic, OpenAI, or Google) and deploy it wherever your infrastructure runs.
When you'd use it
Your team needs quick answers about your architecture without context switching.
During code reviews. A reviewer sees an event being published and wants to know what consumes it. They @mention the bot in the PR channel: "What services consume the PaymentProcessed event?" The bot responds with the current list from your catalog.
During incidents. The payment service is down. Your on-call engineer needs to know what events it produces and what depends on them. They ask the bot in the incident channel. They get the architecture context immediately without leaving the conversation.
Onboarding new developers. Someone joins the team and needs to understand the Orders domain. They post in your dedicated #ask-eventcatalog channel: "List all events in the Orders domain." The bot provides the overview. They follow up with questions about specific events. The bot uses thread context to provide relevant answers.
During planning. Your team is designing a new feature that needs to emit events. They ask the bot what similar events exist and how they're structured. They get examples from your actual schemas, not outdated documentation.
How it works
The bot connects to three things: your Slack workspace, your EventCatalog instance, and your chosen AI provider.
First, you create a Slack app. EventCatalog provides a manifest that configures all the permissions automatically. Generate three credentials (bot token, app token, signing secret) and you're done. The Slack app setup documentation walks through this in detail.
Second, you configure the bot. Point it at your EventCatalog instance, choose your AI provider, and optionally set up auto-reply channels where the bot responds to every message without needing @mentions.
export default {
eventCatalog: {
url: 'https://your-catalog.example.com',
},
ai: {
provider: 'anthropic',
maxSteps: 5,
temperature: 0.4,
},
slack: {
autoReplyChannels: ['C0123456789'], // #ask-eventcatalog
},
};
Third, you deploy it. The bot runs as a long-lived process that maintains a connection to Slack. Deploy it with Docker, Railway, Fly.io, Render, or any cloud provider. Because it uses Socket Mode, you don't need load balancers, SSL certificates, or public endpoints.
The bot automatically handles thread context. Ask a follow-up question in a thread and it remembers what you were discussing. "Tell me about OrderCreated" followed by "What are its consumers?" works naturally.
Custom tools for real-time data
The bot supports custom tools defined in your eventcatalog.chat.js file. This means you can bring live data into conversations.
Query production metrics from Datadog or Prometheus. Check service health and uptime. Look up on-call engineers from PagerDuty. Fetch queue depths from Kafka. Any integration your team needs.
The AI decides when to use custom tools based on the question. Ask "who is on-call for OrderService?" and get live PagerDuty data. Ask "what's the current error rate for PaymentService?" and get real-time metrics.
This turns the bot into a single interface for both documentation and operational data.
Getting started
Prerequisites:
- EventCatalog instance running in SSR mode with configured MCP server
- EventCatalog Scale license (14-day free trial available)
- Slack workspace with app creation permissions
- API key for your chosen AI provider
Quick start:
Clone the repository and install dependencies:
git clone https://github.com/event-catalog/eventcatalog-slack-bot.git
cd eventcatalog-slack-bot
npm install
Create a .env file with your credentials:
EVENTCATALOG_SCALE_LICENSE_KEY=your-license-key
SLACK_BOT_TOKEN=xoxb-...
SLACK_APP_TOKEN=xapp-...
SLACK_SIGNING_SECRET=...
ANTHROPIC_API_KEY=sk-ant-...
Create eventcatalog-bot.config.ts with your configuration, then start the bot:
npm run dev
Invite the bot to a channel and @mention it with a question. That's it.
Full setup instructions, deployment guides, and configuration options are in the Slack integration documentation.
Summary
The EventCatalog Slack Bot brings your architecture documentation into Slack where your team already works. Teams can ask questions, get instant answers grounded in your catalog, and optionally integrate real-time data through custom tools.
This saves time previously spent searching through wikis, reduces context switching during development and incidents, and makes architecture knowledge accessible to everyone on your team.
Learn more:
Get involved: