GitHub

Getting Started

This quick start shows the minimal steps to install the core JCGE packages and run a reference model. For full workflows and modeling guidance, see the Guides.

Prerequisites

  • Julia 1.9+ installed and on your PATH.
  • A clean project environment (recommended).

Install

Create or activate a project before adding packages:

import Pkg
Pkg.activate(".")

Then add the core packages:

import Pkg
Pkg.add([
    "JCGECore",
    "JCGEBlocks",
    "JCGERuntime",
    "JCGECalibrate",
    "JCGEOutput",
])

Run a reference model

using JCGEExamples
result = JCGEExamples.StandardCGE.solve()

Inspect results

The returned result contains the solved model and output artifacts. Typical next steps include inspecting reports and exporting tables via JCGEOutput.

Solver note

For MCP models, PATHSolver is required. See the Imports guide for license setup.

Agent interface

If you want to use JCGE through an MCP-compatible AI assistant, configure the released MCP server in the client, or install JCGEAgentInterface in a dedicated MCP host environment:

import Pkg
Pkg.add("JCGEAgentInterface")

Your model package does not need to depend on or import the agent interface. The AI Agent Interface guide explains the released server, Docker image, and the separate-host pattern used when an agent needs to run a model.

Next steps