MCP.so
Sign In

Nuanced MCP Server

@MattMorgis

About Nuanced MCP Server

A Model Context Protocol (MCP) server that provides call graph analysis capabilities to LLMs through the nuanced library

Basic information

Category

Other

Runtime

python

Transports

stdio

Publisher

MattMorgis

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "nuanced": {
      "command": "uv",
      "args": [
        "--directory",
        "/path/to/nuanced-mcp",
        "run",
        "nuanced_mcp_server.py"
      ]
    }
  }
}

Tools

6

Initialize a code graph for the given repository path

Switch to a different initialized repository

List all initialized repositories

Get the call graph for a specific function

Find all module or file dependencies in the codebase

Analyze the impact of changing a specific function

Overview

What is Nuanced MCP Server?

A Model Context Protocol (MCP) server that provides call graph analysis capabilities to LLMs through the nuanced library. It enables AI assistants to understand code structure by accessing function call relationships and dependencies.

How to use Nuanced MCP Server?

Configure it with Claude Desktop by adding the provided UV-based JSON entry to your claude_desktop_config.json, specifying the path to the server script.

Key features of Nuanced MCP Server

  • Initialize call graphs for Python repositories
  • Explore function call relationships and dependencies
  • Analyze change impact for specific functions
  • Retrieve detailed function information via resources
  • Use pre-built prompts for function and dependency analysis

Use cases of Nuanced MCP Server

  • Give an LLM a structural overview of a codebase before answering questions
  • Assess the impact of modifying a function before making changes
  • Trace dependencies across modules or files in a Python project
  • Automatically generate context-aware code assistance based on call graphs

FAQ from Nuanced MCP Server

What programming languages does it support?

It supports Python repositories for call graph analysis.

How do I configure it with Claude Desktop?

Add the provided JSON configuration using uv to your claude_desktop_config.json, pointing to the server script’s directory.

What tools are available?

Tools include initialize_graph, switch_repository, list_repositories, get_function_call_graph, analyze_dependencies, and analyze_change_impact.

What resources are available?

Resources include graph://summary, graph://repo/{repo_path}/summary, and graph://function/{file_path}/{function_name}.

Comments

More Other MCP servers