Compiler Explorer MCP server

This MCP server connects LLMs to the Compiler Explorer API, allowing them to compile code and analyze optimizations across different compilers and languages. It enables exploration of compiler features, optimization analysis, and detailed assembly comparison without leaving your chat interface.

Installation

To set up the Compiler Explorer MCP server:

1. Clone the repository:

   git clone https://github.com/compiler-explorer/compiler-explorer-mcp.git
   cd compiler-explorer-mcp
   

2. Install the required dependencies:

   npm install
   

3. Start the server:

   npm start
   

By default, the server will run on port 3000. You can modify this by setting the PORT environment variable.

Configuration

API Configuration

You can configure the MCP server by setting environment variables:

PORT=3001 npm start

Connecting to the MCP Server

The MCP server implements the Model Context Protocol, allowing compatible LLMs to connect to it. Configure your LLM tool to use the server URL:

http://localhost:3000/mcp

Usage Examples

Here are some ways to leverage the Compiler Explorer MCP with your LLM:

Compiler Feature Exploration

Ask your LLM questions about compiler features:

• "What is the earliest version of GCC that supports the #embed directive?"
• "Show me how different versions of Clang handle C++20 modules"
• "What optimization flags are available in Clang 12 that weren't in Clang 11?"

Optimization Analysis

Analyze how different optimization levels affect code:

• "What's the assembly difference between -O2 and -O3 for a simple recursive Fibonacci function in GCC 13?"
• "How does Clang's vectorization compare to GCC's for a basic matrix multiplication?"
• "What's the impact of -ffast-math on this floating-point heavy computation?"

Assembly Deep Dives

Explore assembly generation across compilers:

• "What's the most efficient way to implement a population count in x86 assembly across different CPU architectures?"
• "Show me how different compilers optimize a simple string reverse function at -O3"
• "Compare the assembly output of a virtual function call vs a normal function call"

Cross-Language Comparison

Compare implementations across programming languages:

• "Compare the generated assembly for the same algorithm in C++, Rust, and Go"
• "How do exception handling mechanisms differ between C++ and Rust in terms of generated code?"
• "Show me the overhead of Rust's bounds checking compared to unchecked C++ array access"

Troubleshooting

If you encounter issues:

1. Check that the server is running by visiting http://localhost:3000 in your browser
2. Ensure your LLM tool is properly configured to use the MCP endpoint
3. Verify network connectivity if using a remote server

For connection issues, try restarting the server or checking for error messages in the console output.

How to add this MCP server to Cursor

There are two ways to add an MCP server to Cursor. The most common way is to add the server globally in the ~/.cursor/mcp.json file so that it is available in all of your projects.

If you only need the server in a single project, you can add it to the project instead by creating or adding it to the .cursor/mcp.json file.

Adding an MCP server to Cursor globally

To add a global MCP server go to Cursor Settings > MCP and click "Add new global MCP server".

When you click that button the ~/.cursor/mcp.json file will be opened and you can add your server like this:

{
    "mcpServers": {
        "cursor-rules-mcp": {
            "command": "npx",
            "args": [
                "-y",
                "cursor-rules-mcp"
            ]
        }
    }
}

Adding an MCP server to a project

To add an MCP server to a project you can create a new .cursor/mcp.json file or add it to the existing one. This will look exactly the same as the global MCP server example above.

How to use the MCP server

Once the server is installed, you might need to head back to Settings > MCP and click the refresh button.

The Cursor agent will then be able to see the available tools the added MCP server has available and will call them when it needs to.

You can also explictly ask the agent to use the tool by mentioning the tool name and describing what the function does.