STK‑MCP Overview

STK‑MCP is a Model Context Protocol server that bridges Large Language Models (LLMs) and other MCP clients to the powerful digital mission‑engineering platform Ansys/AGI Systems Tool Kit (STK). By exposing STK’s rich simulation and analysis capabilities as a set of MCP tools, developers can invoke complex orbital mechanics, sensor models, or mission planning workflows directly from an AI assistant. This eliminates the need for manual scripting or GUI interaction, allowing LLMs to generate, run, and interpret mission scenarios in a single conversational session.

The server supports both STK Desktop (Windows only) and the cross‑platform STK Engine, automatically selecting the appropriate mode based on the operating system. When launched, STK‑MCP starts a fresh STK instance, keeps it alive for the duration of the MCP session, and shuts it down cleanly when the server stops. This lifecycle management frees developers from juggling separate launch scripts and ensures that every tool call operates on a consistent, isolated environment.

Key capabilities include:

• Tool discovery – a command enumerates all available MCP tools, making it easy to see what actions the server can perform.
• Dynamic tool exposure – STK operations such as creating objects, defining orbits, running propagations, and extracting telemetry are wrapped in MCP tools that can be called by name with JSON payloads.
• OS‑aware deployment – Desktop mode is automatically disabled on non‑Windows platforms, while Engine mode works seamlessly on both Windows and Linux.
• CLI integration – a Typer‑based command line interface provides a familiar developer experience for starting, stopping, and configuring the server.

Typical use cases involve an AI assistant that drafts mission plans: the user asks for a “low‑Earth orbit satellite launch with a 30‑day mission”, and the assistant translates that into a series of MCP tool calls to create launch vehicles, propagate trajectories, and generate coverage reports—all without the user writing Python code. In research or engineering workflows, STK‑MCP enables batch simulation pipelines that can be triggered from LLMs or other automation tools, dramatically accelerating the design‑evaluate‑iterate cycle.

Because STK is a domain‑specific, high‑fidelity simulation engine, having it exposed through MCP gives AI developers unparalleled access to realistic physics and mission analysis. The server’s modular architecture (CLI, MCP logic, STK abstraction layer, and tool definitions) allows extensions to new STK features without touching the core protocol logic. This combination of robustness, flexibility, and ease of integration makes STK‑MCP a standout solution for embedding sophisticated mission engineering into AI‑driven applications.