@inkr/codebase-mcp

An MCP (Model Context Protocol) server designed to provide tools for interacting with codebases, leveraging generative AI capabilities. This server allows AI agents or other MCP clients to ask questions about a specified codebase path.

✨ Features

• Codebase Interaction: Provides an MCP tool (ask_codebase) to query information about codebases.
• Generative AI Integration: Utilizes @google/genai for processing and answering codebase-related questions.
• Stdio Transport: Communicates using the standard MCP stdio transport mechanism.

📚 Table of Contents

• @inkr/codebase-mcp
  • ✨ Features
  • 📚 Table of Contents
  • 🚀 Usage
    • Connecting an MCP Client
    • Environment Variables
    • Tools
      • ask_codebase
  • ⚙️ Development
    • Prerequisites
    • Steps
  • 💻 Technology Stack
  • 📜 License
  • 📧 Contact

🚀 Usage

This project runs as an MCP server. It's typically invoked by an MCP client or controller.

Connecting an MCP Client

Follow the configurations in these files, remember to update env inside with your preference:

• Visual Studio Code: .vscode/mcp.json
• Claude: .mcp.json
• Cursor: .cursor/mcp.json
• Roo Code: .roo/mcp.json

Environment Variables

| Variable | Description | Required | Default | | ---------------- | -------------------------------------- | -------- | --------------------------- | | GEMINI_API_KEY | Your Gemini API key | Yes | | | MODEL | The model to use for codebase analysis | No | gemini-2.0-flash | | CODEBASE_PATH | The path to the codebase | No | (current working directory) |

Tools

Once connected, the client can invoke the tools provided by this server.

ask_codebase

| Argument | Description | Required | Default | | ---------- | ----------------------------------------------------------------------- | -------- | --------------------- | | question | The question to ask about the codebase | Yes | | | path | The path of the file or directory to query, relative to CODEBASE_PATH | No | (Use CODEBASE_PATH) |

⚙️ Development

Prerequisites

• Bun
• Git

Steps

1. Install dependencies:

   bun install

2. Configuration (if necessary):

   • Check if any specific environment variables are required for @google/genai (e.g., API keys). Create a .env file if needed. (Further investigation might be needed to confirm required variables).

3. Run:

   • In IDEs: Reload window and check if the MCP is connected
   • Manually: Run ./run in your terminal

💻 Technology Stack

• Runtime: Bun / Node.js
• Language: TypeScript
• Core Libraries:
  • @modelcontextprotocol/sdk: For MCP server implementation.
  • @google/genai: For generative AI features.
  • Zod: For schema validation.
• Development: @types/bun, TypeScript

📜 License

This project is licensed under the MIT License - see the LICENSE file for details. Copyright (c) 2025 INKR Global

📧 Contact

• Khoa Nguyen @ [email protected]
• INKR Global @ inkr.com