code-intelligence-mcp
Indexes Python codebases using AST and BM25, providing 8 MCP tools for code exploration and search.
README
Code Intelligence MCP Server
A Model Context Protocol (MCP) server that indexes Python codebases and exposes intelligent tools for code exploration and search.
Overview
This server analyzes a Python project at startup, building an in-memory index that includes:
- AST parsing: Extracts classes, functions, methods with signatures and docstrings
- Import tracking: Maps what each file imports
- Usage tracking: Finds where symbols are referenced throughout the codebase
- BM25 search: Full-text keyword search across all code
The indexed data powers 8 MCP tools that let an AI assistant navigate and query codebases semantically.
Quick Start (uvx)
Run directly without installing:
uvx code-intelligence-mcp --root D:\path\to\your\project
Note: On Windows, use double backslashes or forward slashes for paths.
Directory Structure
code_intelligence_mcp/
├── main.py # Entry point - run this to start the server
├── mcp_client.py # Test client that connects to the server
├── pyproject.toml # Package configuration
├── README.md # This file
├── src/
│ ├── __init__.py # Package exports
│ ├── __main__.py # Package entry point (for python -m src)
│ ├── types.py # SymbolInfo, CodeIndex data classes
│ ├── indexer.py # AST parsing and codebase indexing
│ ├── search.py # BM25 search setup and search logic
│ └── tools.py # MCP tool definitions (create_tools, handle_tool)
└── code_intelligence.py # Original monolithic file (kept for reference)
Requirements
- Python 3.12+
uvpackage manager- Dependencies (installed in project virtual environment):
langchain-mcp-adapters>=0.3.0mcp>=1.27.2rank-bm25>=0.2.2
Install Dependencies
From the repo root:
uv add mcp rank-bm25
If you already have them installed, skip this step.
Run the MCP Server
Option 1: uvx (recommended for quick testing)
uvx code-intelligence-mcp --root D:\path\to\your\project
Option 2: Local development
cd code_intelligence_mcp
uv run python main.py --root D:\path\to\your\project
Option 3: python -m src
cd code_intelligence_mcp
uv run python -m src --root D:\path\to\your\project
Note: On Windows, pass the root path with normal backslashes and use
uv run pythoninstead of callinguvdirectly as the subprocess command.
The server will print indexing progress and then wait for MCP requests:
[code-intel] Indexing: D:\path\to\project
[code-intel] Setting up BM25 search index...
[code-intel] Done — 42 files, 156 symbols, 320 BM25 chunks
MCP Tools
1. get_overview
High-level overview of the entire codebase. Call this first to orient yourself.
- Returns: project root, total files/classes/functions, file tree, all symbol names
2. get_file
Get the full source of a specific file by relative path.
- Arguments:
path(string) - relative path to the file - Returns: Full file contents with path header
3. find_symbol
Look up a class or function by exact name.
- Arguments:
name(string) - class or function name - Returns: File, line, docstring, signature, and (for classes) method signatures
4. search_code
Full-text search using BM25 algorithm - good for finding code by keyword or phrase.
- Arguments:
query(string, required) - search keyword/phrasetop_k(integer, default=5) - number of results
- Returns: Top matching code snippets with scores, file, and line context
5. get_imports
Get all imports for a specific file - what modules/names it depends on.
- Arguments:
path(string) - relative path to the file - Returns: List of imported module names
6. find_usages
Find every location in the codebase where a given name is referenced (calls, attribute access, variable use).
- Arguments:
name(string, required) - symbol name to findlimit(integer, default=20) - max results
- Returns: Total usage count and list of {file, line} locations
7. list_classes
List ALL classes in the codebase with their file, line, and method signatures.
- Returns: Array of {name, file, line, docstring, methods} sorted by file/line
8. list_functions
List ALL top-level functions with their file, line, signature, and docstring.
- Returns: Array of {name, signature, file, line, docstring} sorted by file/line
Example Usage
Test with mcp_client.py
import asyncio
from langchain_mcp_adapters.client import MultiServerMCPClient
async def main():
client = MultiServerMCPClient(
{
"project-intel": {
"transport": "stdio",
"command": "uv",
"args": [
"run",
"python",
"main.py",
"--root",
"D:\\path\\to\\your\\project",
],
}
}
)
tools = await client.get_tools()
print(f"Loaded {len(tools)} tools:")
for tool in tools:
print(f" - {tool.name}")
if __name__ == "__main__":
asyncio.run(main())
Run with:
uv run python mcp_client.py
Install from PyPI
After publishing, users can install and run:
# Install globally
uv pip install code-intelligence-mcp
# Or run directly with uvx (recommended)
uvx code-intelligence-mcp --root D:\path\to\your\project
Troubleshooting
mcp.shared.exceptions.McpError: Connection closedusually means the server subprocess failed to start correctly.- Verify the
uv run pythoncommand works manually before using it insidemcp_client.py. - Ensure the path passed to
--rootis valid and accessible. - If you get import errors, make sure you're running from the project directory with the correct virtual environment activated.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.