kannajune/mcp-architect

Overview

A Python tool that serves architectural understanding of a codebase to any AI assistant via the Model Context Protocol. It runs locally and privately with zero configuration, providing context about code structure without sending code to external services.

Best for

Best for
Developers who want to give local AI assistants deep codebase understanding without sending code to external services

Use cases

• Give an AI assistant a high-level map of code dependencies and modules
• Enable AI tools to answer architecture questions about a private codebase
• Integrate codebase context into local AI workflows without cloud uploads

Notes

A Python tool that serves architectural understanding of a codebase to any AI assistant via the Model Context Protocol. It runs locally and privately with zero configuration, providing context about code structure without sending code to external services.

1 stars on GitHub. Last updated 2026-06-18. Licensed MIT.

Use cases

• Give an AI assistant a high-level map of code dependencies and modules
• Enable AI tools to answer architecture questions about a private codebase
• Integrate codebase context into local AI workflows without cloud uploads

Pros

• Runs entirely local and private, keeping code on the machine
• Zero-config setup reduces friction for quick adoption
• Provides architectural context beyond raw file contents

Cons

• Very early stage with only 1 GitHub star and limited community validation
• Requires an AI assistant that supports the Model Context Protocol
• May lack robustness or documentation for complex codebases

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.