VSCodeMCP
About
A simple, robust, mcp inference server for groq. Tailored specifically for vscode copilot mcp access.
Explore
- 🛠️ Easy Tool Definition: Define MCP tools with simple Python functions
- 📊 VS Code Configuration: Generate VS Code config files automatically
- 🔗 FastMCP Integration: Seamless integration with the FastMCP library
- 🧠 Groq LLM Support: Use Groq's powerful LLMs in your MCP tools
- 📱 URL Generation: Generate VS Code URLs for easy server installation
- 🔐 Secure Input Handling: Manage API keys and sensitive inputs securely
- 📝 Auto-Generated Scripts: Automatic generation of server scripts
Setting up with Highlight
This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:
- Download and install Highlight from highlightai.com/download
- Navigate to the plugins tab and select "Add Custom Plugin"
-
Configure the plugin with the settings below
Plugin Name
VSCodeMCPCommand (node, npx, python, etc.)Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.
- Enable "Start Automatically" if you want the plugin to start when Highlight launches
From the repository
pip install fastmcp
The module is designed to work seamlessly with FastMCP for implementing custom tool functionality:
from fastmcp import FastMCP
from vscode_mcp import VSCodeMCP, ServerType
python -m vscode_mcp install path/to/server.py --user
```
server.add_tool(
name="hello_world",
description="Responds with a friendly greeting"
)
server.add_tool(
name="fetch_data",
description="Fetches data from an external API"
)
server.add_chat_tool(
name="ask_groq",
description="Ask Groq a question",
system_prompt="You are a helpful, knowledgeable assistant.",
model=GroqModel.LLAMA_3_70B
)
server.add_code_tool(
name="code_with_groq",
description="Generate code using Groq",
system_prompt="You are an expert programmer.",
model=GroqModel.DEEPSEEK_DISTILL_LLAMA_70B # Using a specialized code model
)
mcp = FastMCP("MyMCPServer")
@mcp.tool()
def hello_world(name="World"):
"""
Responds with a friendly greeting
"""
return {
"content": f"Hello, {name}!"
}
@mcp.tool()
def search_files(pattern, directory="."):
"""
Search for files matching a pattern
"""
import glob
import os
files = glob.glob(os.path.join(directory, pattern))
return {
"content": f"Found {len(files)} files: {', '.join(files)}"
}
Claude Desktop / Cursor
Paste into your MCP client config file to install this server.
{
"mcpServers": {
"vscodemcp": {
"groq-mcp": {
"command": "python",
"args": [
"-m",
"vscode_mcp",
"create",
"My MCP Server",
"--tool",
"hello_world",
"Responds with a greeting",
"--install"
]
}
}
}
}McpServers
{
"groq-mcp": {
"command": "python",
"args": [
"-m",
"vscode_mcp",
"create",
"My MCP Server",
"--tool",
"hello_world",
"Responds with a greeting",
"--install"
]
}
}TODO RETIRE OR TURN INTO JUST A VSCODE SCRIPT IF THE OTHER GROQ-MCP-SERVER FOR DOESNT WORK OUT THE BOX WITH VSCODE COPILOT THEN WE NEED TO FIGURE OUT WHAT THAT ISSUE IS
A Python module for creating and managing MCP (Model Context Protocol) servers for Visual Studio Code's Copilot agent feature, with optional Groq LLM integration.
Overview
The VSCodeMCP module makes it easy to:
1. Define MCP tools, resources, and prompts for VS Code's Copilot agent
2. Generate configuration files for VS Code
3. Install servers using VS Code's command line interface
4. Create shareable configuration files for your workspace
5. Integrate with Groq LLMs for powerful AI-driven tools
Features
- 🛠️ Easy Tool Definition: Define MCP tools with simple Python functions
- 📊 VS Code Configuration: Generate VS Code config files automatically
- 🔗 FastMCP Integration: Seamless integration with the FastMCP library
- 🧠 Groq LLM Support: Use Groq's powerful LLMs in your MCP tools
- 📱 URL Generation: Generate VS Code URLs for easy server installation
- 🔐 Secure Input Handling: Manage API keys and sensitive inputs securely
- 📝 Auto-Generated Scripts: Automatic generation of server scripts
Installation
```bash
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
