PeopleInSpace

Kotlin Multiplatform sample with SwiftUI, Jetpack Compose, Compose for Wear, Compose for Desktop, and Compose for Web clients along with Ktor backend.

Installation

Run this command in your terminal to add the MCP server to Claude Code.

Run in terminal:

View docs

Command

claude mcp add --transport stdio joreilly-peopleinspace java -jar /path/to/PeopleInSpace/mcp-server/build/libs/serverAll.jar --stdio

How to use

PeopleInSpace includes an MCP server that exposes a tool endpoint for listing the people currently in space using the shared Kotlin Multiplatform codebase. The MCP server is built from the mcp-server module and is intended to be run alongside the rest of the Kotlin/KMP project. To use it in integrations (for example, Claude Desktop), you first build a standalone JAR with the shadowJar task, then start the server with the provided stdio options. The MCP server can be queried via the standard MCP tools interface and can be plugged into client tooling that expects an MCP endpoint returning the current astronauts in space. You can reference the example configuration in the repository to wire the server into your MCP environment or Claude integration.

Once running, you’ll typically interact with the server by starting the JAR and routing input/output through stdio, enabling you to feed commands and receive structured responses that describe the people in space. This makes it easy to connect the MCP server to external clients that need real-time spaceperson data without reimplementing the fetch/parse logic in each consumer.

How to install

Prerequisites:

Java JDK (recommended 17+) installed and available on your PATH
Gradle installed or use the Gradle wrapper included in the project

Steps:

Clone the repository or navigate to the PeopleInSpace project root.
Build the MCP server using the shadowJar task for the backend, which produces a serverAll.jar containing the MCP server:

cd PeopleInSpace
./gradlew :mcp-server:shadowJar

Locate the built JAR (e.g., mcp-server/build/libs/serverAll.jar).
Run the MCP server via Java, using stdio mode as shown in the example configuration:

java -jar mcp-server/build/libs/serverAll.jar --stdio

If you need to integrate with an MCP client configuration, create an mcp.yaml or JSON similar to the example in the repository, pointing to the path of serverAll.jar and enabling stdio I/O.

Optional: If you are also building the full Kotlin Multiplatform project, you can build other modules (backend, compose desktop, etc.) with the standard Gradle commands used in the project.

Additional notes

Tips and common notes:

Ensure the path to serverAll.jar in your mcp configuration matches your local build output.
The MCP server uses the Kotlin MCP SDK. This means you’ll typically interact via the MCP tool interface rather than a full HTTP REST endpoint.
If you modify the backend or shared KMP code, re-run the shadowJar task to refresh serverAll.jar.
For debugging, you can run the server in stdio mode and test with sample MCP clients or with the Claude Desktop integration example provided in the README.
Memory and environment: allocate sufficient heap if the MCP server pulls in large data sets or caches; tune JVM options as needed for your deployment environment.