Poop MCP Server

The Poop MCP Server is a lightweight, Java‑based Model Context Protocol (MCP) implementation that enables AI assistants to perform two distinct external tasks through a single, unified interface: manipulating objects in a 3D scene and retrieving live cryptocurrency prices. By exposing these services as MCP tools, the server removes the friction of integrating disparate APIs into conversational agents, allowing developers to focus on crafting intelligent user experiences rather than plumbing.

At its core, the server offers two command families. The 3D Scene Command Service accepts movement directives—“move” commands with a target direction (front, back, left, right) and a distance value—processes them against an internal scene graph, and returns a structured JSON payload describing the updated state. This is ideal for applications such as virtual reality walkthroughs, interactive simulations, or game‑like storytelling where an AI agent must guide a user through spatial navigation. The Cryptocurrency Price Service queries the CoinGecko API for real‑time price data on any supported coin (e.g., bitcoin, ethereum) and surfaces the information in a consistent JSON format. This makes it straightforward to embed up‑to‑date market data into finance chatbots, portfolio trackers, or investment advisory tools.

Key capabilities include:

• Standardized MCP contract: Tools are exposed via and annotations, ensuring that any MCP‑compliant client can discover and invoke them without custom adapters.
• Spring AI integration: Built on Spring AI’s tool framework, the server can be dropped into existing Spring Boot applications, leveraging dependency injection and configuration management.
• Docker readiness: A single Dockerfile allows rapid deployment in cloud or edge environments, with environment variables for API keys and profile selection.
• Multi‑profile support: Separate , , and configurations let developers test locally while keeping production secrets isolated.

Typical use cases span from interactive educational platforms where an AI tutor manipulates 3D models to illustrate concepts, to financial assistants that pull live market data on demand. In a workflow, an AI model receives user intent, translates it into an MCP tool call (e.g., “move the camera to the left by 5 units”), and receives a JSON response that can be rendered in a UI or fed back into the conversation. The same mechanism fetches cryptocurrency prices, enabling real‑time dialogue about market trends.

What sets Poop MCP Server apart is its dual focus on spatial manipulation and real‑time data retrieval, combined with a minimal footprint. Developers can quickly bootstrap the server, integrate it into their AI stack, and expose powerful external capabilities to conversational agents without wrestling with low‑level HTTP clients or state management.