mcp-remote-call-ping-pong
mcp-remote-call-ping-pong is a tool designed for bidirectional communication between remote systems. It is particularly useful for testing and debugging API calls. The tool features a simple HTML interface, making it easy for users to operate. It allows for visual confirmation of data transmission, making it highly beneficial for developers and testers.
GitHub Stars
1
User Rating
Not Rated
Favorites
0
Views
2
Forks
2
Issues
1
🏓 MCP Ping-Pong Server by Remote Call
An experimental and educational Ping-Pong server demonstrating MCP (Model Context Protocol) calls via FastAPI.
Features
- ✅ FastAPI/FastMCP backend for remote MCP calls through API endpoints or SSE
- 🔄 MCP integration for command handling
- 🔐 Thread-safe session management
Quick Start
Installation
To install the dependencies, run:
poetry install
1. Call MCP Tool through API
Run the Server
Start the FastAPI server with:
python mcp-api-server.py
The server will start at http://localhost:8080.
UI
Open mcp-api-client.html in a browser to interact with the UI.
API Endpoints
GET /ping-pong?prompt_name=<prompt_name>: Retrieves the specified prompt.POST /ping-pong: Invokes MCP tool commands (ping,pong,count).
2. Call MCP Tool through SSE transport
The following example demonstrates using Server-Sent Events (SSE) for communication with the MCP server.
Run the SSE Server
python mcp-sse-server.py
Run the SSE Client
python mcp-sse-client.py
Sample Output
Type a command (e.g., 'ping', 'pong', 'count') or 'exit' to quit:
>>> ping
Sending command: ping
Result: pong
>>> pong
Sending command: pong
Result: ping
>>> count
Sending command: count
Result: 2
References
📄 License
MIT
MCP Playground is a Streamlit-based interface that allows users to interact with large language models while seamlessly integrating external Multi-Server Command Protocol (MCP) tools. It enables the deployment of multiple FastMCP servers managed via Docker Compose, creating a provider-agnostic client using LangChain and LangGraph.