What is EventCatlaog Chat?

Added in eventcatalog@2.24.2

License: Dual-license

EventCatalog Chat is designed to save you and your team time. It's a chat interface that allows you to ask questions about your architecture. You can ask questions about your architecture, and it will return the most relevant information to you.

Your development teams, product managers, and architects can all benefit from EventCatalog Chat. Finding information about your architecture can be a time consuming activity. EventCatalog Chat can help you save time and get the information you need.

Example use cases:

• Quickly get the messages related to a feature
  • "I have a new feature that does (X), what events in our architecture are related to this?"
• Dive deeper into your architecture
  • "I want to know more about the "Order" service, what events are produced and consumed?"
• Find resources, ask questions and dive deeper into your catalog
  • "I want to know more about the "Payment" domain, what services are related to it?"

In the example below, we are asking EventCatalog Chat to write a JSON schema for a given message following company best practices (with custom prompts).

With EventCatalog Chat you can bring your own keys (BYOK) for your own models, or use the default model (browser based). If you want better results, you can use your own model.

"EventCatalog is tech agnostic"

Remember, EventCatalog can be generated from any broker, API or system. So you can create your catalog powered by your broker, OpenAPI or AsyncAPI specs and then ask questions about your architecture.

Features

• Bring your own keys (BYOK): Bring your own keys for your own models, currently only supports OpenAPI.
• In-browser models: You can opt into use our browser based models. If you choose to use the browser based models, your data is never sent to a server and is private.
• Chat history: Access and manage your conversation history.
• Code generation: Generate code from your messages, services, and domains.
• Bring your own prompts: Bring your own prompts to customize the chat experience. Set your own company best practices. Get better results for all your teams using EventCatalog Chat.

Browser support

By default, WebGPU is enabled and supported in both Chrome and Edge. However, it is possible to enable it in Firefox and Firefox Nightly. Check the browser compatibility for more information.

Hardware requirements

To run the models efficiently, you'll need a GPU with enough memory. 7B models require a GPU with about 6GB memory whilst 3B models require around 3GB.

Smaller models might not be able to process file embeddings as efficient as larger ones.

You can configure which model to use in the eventcatalog.config.js file.

How does it work?

EventCatalog Chat uses a local-first approach to keep your data secure and local. Models are downloaded to your browser, and your catalog is used as a source of reference. The model will use vector embeddings to find the most relevant information in your catalog.

Quick setup

1. First you create the embeddings for your catalog, this is done using the @eventcatalog/generator-ai plugin.
2. You configure your eventcatalog.config.js file to enable the AI assistant.
3. You can now start chatting with your catalog.
   • The first time you start the chat, the model will be downloaded to your browser (this can take a couple of minutes)
   • Subsequent chats will use the model that is cached in your browser.

Commercial use

During beta, this feature is free to use.

Once released this feature will be governed by a dual-license. To ensure the sustainability of the project.

If you would like to use this feature in a commercial product, you can purchase a license by emailing us at hello@eventcatalog.dev.

Have any questions? You can email us at hello@eventcatalog.dev.