Demo: Confluent Intelligence - Real-Time Context Engine

Añadido: 2026-05-23

[00:00:00]AI is only as good as its context, but in most enterprises that context is either stale, fragmented across systems, or it's exposed as raw streams that are hard for AI apps to use effectively.

[00:00:12]That's what Confluent Real-time Context Engine solves.

[00:00:15]Real-time Context Engine continuously serves fresh structured context to any AI system through MCP as a fully managed service in Confluent Cloud. In this demo, we'll walk through a quick example of Real-time Context Engine in action.

[00:00:29]So, our use case here is pit management for an F1 race.

[00:00:32]I'm here in the Confluent Cloud console and I have several useful topics, but the one that I want to expose for Real-time Context here is race standings. I'm going to scroll over. I'm going to enable Real-time Context Engine on this topic. Now, we need to configure the MCP server and AI tool of our choice. In this case, I'm going to use Cursor. So, here in Cursor, I'm going to go to settings. I'm going to search for MCP, and then I'll select tools and MCPs, and then add custom MCP. This opens a JSON file that I can edit. So, I'm going to paste in all of the details for our Confluent Cloud MCP server.

[00:01:05]Again, fully managed by Confluent Cloud here. You just need the URL and your auth token, and then we'll save this file. And then when I go back to Cursor settings, you can see here there's a green dot on my MCP server, and then I can also see the three tools that are exposed by the MCP server. Get metadata, list topics, and query data. So, let's give all three of these a try. Now, the first question I'm going to ask my agent is what topics are available Real-time Context Engine?

[00:01:33]We will allow list this MCP run, and then you can see our race standings topic is listed here. Next, I want to understand the schema of this topic. So, I'm going to say give me the schema for the race standings topic.

[00:01:45]And here we have full detail of the schema. Now, the final question I'm going to ask is what car is currently in position one? This will actually allow us to query data on this topic and get a response here in our agent. And we can see here that the car in position one is number 44.

[00:02:01]So there you have it. With real-time context engine from Confluent, you get fresh real-time data exposed to your AI tools from MCP.

[00:02:09]All we had to do was enable real-time context engine on our topic, and then plug in the MCP server details in the tool of our choice. If you want to dig deeper into real-time context engine and try it for yourself, check out our latest blog post as well as our Confluent Cloud documentation that details how to get started with real-time context engine.