Mibayy/token-savior

What it does

Token Savior is a structural code intelligence MCP server designed for AI-assisted development. It allows developers and AI systems to navigate and edit code using semantic symbol names instead of raw text, enabling intelligent code comprehension without excessive token consumption. The tool provides advanced code analysis capabilities that significantly reduce token usage in AI interactions while improving development efficiency.

Key features

• Symbol lookup and navigation by semantic names
• Dependency graph visualization and analysis
• Call chain tracking and traversal
• Dead code detection and removal
• Structural diff comparisons
• 87% token reduction compared to traditional text-based approaches
• Multi-platform support (Python, macOS, Windows, Linux)

How to set up

Token Savior is implemented as an MCP server compatible with platforms supporting the Model Context Protocol. Installation typically involves cloning the GitHub repository (Mibayy/tokensavior) and configuring it as an MCP server within your AI development environment. Detailed setup instructions are available in the project's GitHub documentation.