Use This MCP server To

Automate code reviews for GitHub pull requests Provide detailed feedback on code quality and style Detect potential issues in pull request code automatically Integrate AI-driven code review into CI/CD pipelines Log and debug code review feedback systematically Support review workflows for public and private repos Enhance developer productivity with automated PR analysis

README

Code Reviewer MCP Server

A GitHub Pull Request Code Review server using Model Context Protocol

This MCP server provides automated code review capabilities for GitHub pull requests by integrating with OpenRouter's AI models.

Features

Automated code reviews for GitHub pull requests
Detailed feedback on code quality, style and potential issues
Integration with OpenRouter's AI models (currently using Gemini 2.5 Pro)
Comprehensive logging for debugging

Requirements

A GitHub Personal Access Token (PAT) with repo scope for private repositories or public_repo scope for public repositories. This PAT is required to fetch pull request data.
An OpenRouter API Key.

Current Limitations

⚠️ Important: Currently this server only works with:

Open pull requests (not draft PRs)

Installation

Clone this repository:

git clone https://github.com/joaomj/code-reviewer-mcp.git
cd code-reviewer-mcp

Install dependencies:

npm install

Set required environment variables:

export GITHUB_PAT=your_github_personal_access_token
export OPENROUTER_API_KEY=your_openrouter_api_key

Build the server:

npm run build

Usage

To start the MCP server:

node build/index.js

The server will be available to any MCP-compatible client (like Claude Desktop).

Configuration

Add to your MCP client configuration:

{
  "mcpServers": {
    "code-reviewer-mcp": {
      "command": "node",
      "args": ["/path/to/code-reviewer-mcp/build/index.js"],
      "env": {
        "GITHUB_PAT": "your_github_token",
        "OPENROUTER_API_KEY": "your_openrouter_key"
      }
    }
  }
}

Development

For development with auto-rebuild:

npm run watch

To run tests:

npm test

Debugging

Use the built-in logging which outputs to stderr with detailed information about:

GitHub API requests
OpenRouter API calls
Processing steps
Any errors encountered

For more advanced debugging, use the MCP Inspector:

npm run inspector

Next Steps

Planned improvements and features:

Draft pull request analysis
Customizable review templates
Support for multiple AI models (OpenAI, Claude, etc.)
File-specific feedback
Rate limiting and caching

Contributing

Suggestions are welcome. Please open an issue first to discuss what you would like to change.

code-reviewer-mcp FAQ

What permissions are required for the GitHub token?

A GitHub Personal Access Token with 'repo' scope for private repos or 'public_repo' scope for public repos is required to fetch pull request data.

Which AI models does this server use for code review?

It currently uses OpenRouter's Gemini 2.5 Pro model for generating code review feedback.

Can this server review draft pull requests?

No, it currently only supports open pull requests, not draft PRs.

How do I install and set up the code-reviewer-mcp server?

Clone the repository, install dependencies with npm, and configure GitHub PAT and OpenRouter API keys as per instructions.

Does the server provide logs for debugging?

Yes, it includes comprehensive logging to help debug and track code review operations.

Can this server be integrated into existing CI/CD workflows?

Yes, it can be integrated to automate code review as part of continuous integration pipelines.

Is this server limited to GitHub repositories only?

Yes, it is designed specifically for GitHub pull request code reviews.

What are the API key requirements?

An OpenRouter API key is required to access the AI models used for code review.