Implementing a MCP server in Quarkus

[giscus[bot]]

Implementing a MCP server in Quarkus

Quarkus: Supersonic Subatomic Java

https://quarkus.io/blog/mcp-server/

[adietish]

Hey Max,

thanks for the nice writeup. Enjoyed reading about MCP.

I failed to run the server using your instructions, step by step:

java -jar target/weather-1.0.0-SNAPSHOT-runner.jar.

The only artifact that was built was target/weather-1.0.0-SNAPSHOT.jar which is not runnable (no main manifest attribute, in target/weather-1.0.0-SNAPSHOT.jar). Ran it using ./mvnw compile quarkus:dev instead.

On the other hand, using the native profile ./mvnw install -Pnative created a executable binary weather-1.0.0-SNAPSHOT-runner that I could run.

What did I miss?

[jcgay]

I had the same issue, I have replaced quarkus.package.uber-jar=true by quarkus.package.jar.type=uber-jar and it seems ok 🤞

[felipenova28]

Hey Max,

When I start my client service with mvn quarkus:dev it gives an error saying that it cannot find the jbang command to run the weather mcp

[jabhijeet]

What is the weather forecast and any alerts for New York?

How a question like this be answered? Do i need to write another api for mixing the commands or will mcp server understand it?

[Alex-abro]

Is there a way to create Tools without annotations? I want to create some Tools dynamically

[Alex-abro]

Update: I think, i found it by my self with ToolManager. See https://docs.quarkiverse.io/quarkus-mcp-server/dev/index.html# Programmatic API

[AlexWareR]

"Excellent article! I appreciated that you covered a complete end-to-end MCP server example with Quarkus from creating the projects to building the weather-forecasting tools to wiring them into the MCP protocol. I think you did a great job of describing core concepts like tools, prompts, and resources within the Java/Quarkus ecosystem that kept the articles easy to follow, and tangible.

I also recently read a complementary piece on https://mobisoftinfotech.com/resources/blog/ai-development/develop-use-mcp-server-ai-agents-maven-guide -which dives deeper into the implementation side especially using Maven and building agent workflows. The focus on MCP server configuration there perfectly complements your tutorial, giving readers a complete picture of how to set up and optimize MCP servers for AI agents."