
mcp transport scratch
一个基于WebSocket和HTTP的MCP传输层实现,包含服务器和客户端代码。
Repository Info
About This Server
一个基于WebSocket和HTTP的MCP传输层实现,包含服务器和客户端代码。
Model Context Protocol (MCP) - This server can be integrated with AI applications to provide additional context and capabilities, enabling enhanced AI interactions and functionality.
Documentation
MCP Transport Scratch
This is a repo containing my tinkering with websocket and HTTP based MCP transport.
It contains a simple MCP server, that I wire up to websockets and http. Custom transport implementations are in transport.http and transport.ws.
I also have a test suite that makes some requests on the MCP server on each transport. To run the suite you need to run some stuff
Setup
uv sync
source ./.venv/bin/activate
export PYTHONPATH=src/
Server
uvicorn server.main:app
Client
python src/client/mcp_client.py
Example output
Starting in-memory client test
Client session started
Client initialized
Tool result: meta=None nextCursor=None tools=[Tool(name='echo_tool', description='Echoes the input', inputSchema={'properties': {'data': {'title': 'Data', 'type': 'string'}}, 'required': ['data'], 'title': 'echo_toolArguments', 'type': 'object'})]
Prompts result: meta=None nextCursor=None prompts=[Prompt(name='test_prompt', description='', arguments=[])]
Resources result: meta=None nextCursor=None resources=[Resource(uri=AnyUrl('test://test_resource'), name='test_resource', description='Test resource', mimeType='text/plain', size=None, annotations=None)]
Call tool result: meta=None content=[TextContent(type='text', text='Hello, world!', annotations=None)] isError=False
Call prompt result: meta=None description=None messages=[PromptMessage(role='user', content=TextContent(type='text', text='test prompt result', annotations=None))]
Read resource result: meta=None contents=[TextResourceContents(uri=AnyUrl('test://test_resource'), mimeType='text/plain', text='test resource data!')]
Memory test completed
Starting websocket client test
Connected to websocket
Client session started
Client initialized
Tool result: meta=None nextCursor=None tools=[Tool(name='echo_tool', description='Echoes the input', inputSchema={'properties': {'data': {'title': 'Data', 'type': 'string'}}, 'required': ['data'], 'title': 'echo_toolArguments', 'type': 'object'})]
Prompts result: meta=None nextCursor=None prompts=[Prompt(name='test_prompt', description='', arguments=[])]
Resources result: meta=None nextCursor=None resources=[Resource(uri=AnyUrl('test://test_resource'), name='test_resource', description='Test resource', mimeType='text/plain', size=None, annotations=None)]
Call tool result: meta=None content=[TextContent(type='text', text='Hello, world!', annotations=None)] isError=False
Call prompt result: meta=None description=None messages=[PromptMessage(role='user', content=TextContent(type='text', text='test prompt result', annotations=None))]
Read resource result: meta=None contents=[TextResourceContents(uri=AnyUrl('test://test_resource'), mimeType='text/plain', text='test resource data!')]
Websocket test completed
Starting HTTP client test
Connected to HTTP
Client session started
Client initialized
HTTP Request: POST http://127.0.0.1:8000/mcp/http "HTTP/1.1 200 OK"
Tool result: meta=None nextCursor=None tools=[Tool(name='echo_tool', description='Echoes the input', inputSchema={'properties': {'data': {'title': 'Data', 'type': 'string'}}, 'required': ['data'], 'title': 'echo_toolArguments', 'type': 'object'})] meta=None
HTTP Request: POST http://127.0.0.1:8000/mcp/http "HTTP/1.1 200 OK"
Prompts result: meta=None nextCursor=None prompts=[Prompt(name='test_prompt', description='', arguments=[])] meta=None
HTTP Request: POST http://127.0.0.1:8000/mcp/http "HTTP/1.1 200 OK"
Resources result: meta=None nextCursor=None resources=[Resource(uri=AnyUrl('test://test_resource'), name='test_resource', description='Test resource', mimeType='text/plain', size=None, annotations=None)] meta=None
HTTP Request: POST http://127.0.0.1:8000/mcp/http "HTTP/1.1 200 OK"
Call tool result: meta=None content=[TextContent(type='text', text='Hello, world!', annotations=None)] isError=False meta=None
HTTP Request: POST http://127.0.0.1:8000/mcp/http "HTTP/1.1 200 OK"
Call prompt result: meta=None description=None messages=[PromptMessage(role='user', content=TextContent(type='text', text='test prompt result', annotations=None))] meta=None
HTTP Request: POST http://127.0.0.1:8000/mcp/http "HTTP/1.1 200 OK"
Read resource result: meta=None contents=[TextResourceContents(uri=AnyUrl('test://test_resource'), mimeType='text/plain', text='test resource data!')] meta=None
HTTP test completed
Quick Start
Clone the repository
git clone https://github.com/jerome3o/mcp-transport-scratchInstall dependencies
cd mcp-transport-scratch
npm installFollow the documentation
Check the repository's README.md file for specific installation and usage instructions.
Repository Details
Recommended MCP Servers
Discord MCP
Enable AI assistants to seamlessly interact with Discord servers, channels, and messages.
Knit MCP
Connect AI agents to 200+ SaaS applications and automate workflows.
Apify MCP Server
Deploy and interact with Apify actors for web scraping and data extraction.
BrowserStack MCP
BrowserStack MCP Server for automated testing across multiple browsers.
Zapier MCP
A Zapier server that provides automation capabilities for various apps.