Set up an MCP Server

Use this topic to set up and configure Model Context Protocol (MCP) connectivity for PPM Aviator.

Start the PPM server

Start the PPM server using your standard process.

Configure the MCP server

The MCP jar comes with the installation and is located under <PPM_HOME>/MCP.

In the same folder, update the application.yaml file with the required details:

server:
  port: <Optional. If this is set to anything other than 8082, update server.conf; otherwise, no update is required.>

ppm:
  baseurl: <IP address of baseURL>

Optionally update the MCP server URL in configuration

Open server.conf.

Update the following property:

com.kintana.core.server.MCP_SERVER_URL = http://localhost:8082/mcp

Start the MCP server

Run the following command from the MCP folder:

java -jar <path to jar file\ppm-mcp-server-0.0.1.jar> --spring.config.location=<path to yaml file\application.yaml>

For example: java -jar c:\MCP1\ppm-mcp-server-0.0.1.jar --spring.config.location=c:\MCP1\application.yaml

Note:

MCP server is required only when using PPM Aviator with latest AI enablement.
For existing/old AI features, MCP server is not required.
Ensure that application.yaml is in the same folder as the MCP jar. If it is in a different location, specify the full path to application.yaml in the MCP jar startup command.

Configure LLM Parameters

To configure the selected LLM provider:

Navigate to Administration > Integrations > AI Model Config.
For each supported provider (for example, Google Vertex and OpenAI):
1. Enter required model parameters.
2. Upload JSON (for Gemini) OR provide API key (for OpenAI).
For example: