Code Context MCP Server

Make any LLM a codebase expert instantly. This MCP server provides deep code intelligence through semantic search, architecture mapping, security analysis, and smart context that fits perfectly in token windows. 90% more efficient than sending raw files. Perfect for Claude, GPT, and any LLM-powered IDE.

Installation

git clone https://github.com/RaheesAhmed/code-context-mcp.git
cd code-context-mcp
uv sync

Usage

# Development mode with inspector
uv run mcp dev src/server.py

Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "code-context": {
      "command": "uv",
      "args": ["run", "--directory", "/path/to/code-context-mcp", "mcp", "run", "src/server.py"]
    }
  }
}

Tools (16 Total)

Core Tools

Tool	Description
get_repo_map	Condensed map of all symbols in entire codebase
get_file_context	File content with related imports
search_symbols	Find function/class definitions
get_dependencies	Import relationships for a file
get_project_stats	File count, lines, language breakdown
read_file	Read file with optional line range

Advanced Search

Tool	Description
find_usages	Find ALL places where a symbol is used
smart_context	Auto-find relevant files for a question
semantic_search	Search code by meaning, not keywords

Deep Analysis

Tool	Description
get_call_graph	Trace callers/callees with Mermaid diagram
get_architecture	Auto-generate project layer diagram
analyze_patterns	Detect security, performance, quality issues

Optimization

Tool	Description
get_compressed_context	Token-efficient multi-file context
analyze_change_impact	What breaks when you change a file
get_recent_changes	Git history - recently modified files
trace_code_flow	Step-by-step execution path tracing

Examples

# See entire codebase structure
get_repo_map(project_path="/path/to/project")

# Find all usages of a function
find_usages(project_path="/path/to/project", symbol="authenticate")

# Auto-find relevant code for a question
smart_context(project_path="/path/to/project", question="how does auth work?")

# Build call graph with visualization
get_call_graph(project_path="/path/to/project", function_name="main")

# Analyze before making changes
analyze_change_impact(project_path="/path/to/project", file_path="src/core.py")

# Trace execution flow
trace_code_flow(project_path="/path/to/project", entry_point="handleRequest")

Requirements

• Python 3.10+
• uv

License

MIT License - see LICENSE

Author

Rahees Ahmed