Code Intelligence

Provides AI-powered UI component and utility method recommendations via MCP with local stdio startup.

typescript

GitHub Stars

typescript

Language

4 months ago

First Indexed

3 weeks ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "lyw405-code-intelligence-mcp": {
      "command": "/bin/zsh",
      "args": [
        "/path/to/code-intelligence-mcp/scripts/start.sh"
      ],
      "env": {
        "CI_MCP_UTILS": "PLACEHOLDER",
        "CI_MCP_CONFIG": "PLACEHOLDER",
        "CI_MCP_DATA_DIR": "PLACEHOLDER",
        "CI_MCP_COMPONENTS": "PLACEHOLDER"
      }
    }
  }
}

Code Intelligence MCP is an intelligent code suggestion service that analyzes your requirements and recommends private UI components and utility methods for rapid, high-quality development within AI IDEs. It exports structured prompts and implementation guidance to help you compose code with reusable building blocks.

How to use

You connect your MCP client to the Code Intelligence MCP server to get AI-powered recommendations for UI components and utility methods. Provide a clear description of what you want to build, and the server will analyze requirements, identify relevant private components and utilities, and generate ready-to-use prompts, imports, and step-by-step implementation guidance.

How to install

Prerequisites: Node.js is required to run MCP tools and servers. You should also have a package manager available, such as npm, pnpm, or npx.

# Ensure Node.js is installed (check version)
node -v
npm -v
pnpm -v

Choose an installation mode that matches your workflow. A stdio-based local server is provided for direct usage from your environment. The server is started via a shell script executed by a runtime command.
Start the MCP server using the explicit startup command. The following example shows the runtime command and arguments as provided for a local setup.

/bin/zsh /path/to/code-intelligence-mcp/scripts/start.sh

Additional setup notes

Environment variables control where the MCP loads its configuration data. The main variable shown is CI_MCP_DATA_DIR, which should point to a directory containing your configuration files. You can also specify paths directly via CI_MCP_CONFIG, CI_MCP_COMPONENTS, and CI_MCP_UTILS.

Notes on configuration and startup

The server expects a configuration directory that provides three data files: model configuration, UI component knowledge base, and utility method knowledge base. The startup command relies on a runtime shell to execute the startup script which launches the MCP service.

Security and maintenance

Keep API keys and private knowledge bases secure. Do not commit sensitive configuration to version control. Regularly update dependencies, review knowledge base entries for accuracy, and monitor usage to ensure prompts remain aligned with your project’s standards.

Troubleshooting

If the MCP service fails to start, verify that the startup script is executable, the specified data directory exists, and environment variables are correctly set. Check logs for errors related to file access, missing configuration files, or permissions.

Available tools

Suggest UI component

Intelligently analyzes UI development requirements and recommends the most suitable private components, returning an optimized prompt and an implementation guide.

Query component

Returns detailed information about a specific UI component, including description, props, events, slots, and usage.

Suggest utilities

Intelligently analyzes logic requirements and recommends reusable utility methods, with an optimized prompt and implementation guide.

Query utility

Returns detailed information about a specific utility method, including parameters, return type, and usage.

Built by

VeilStrat

AI signals for GTM teams