mcpserver.so LogoMCPFly
Submit
Code Review Server logo

Code Review Server

by crazyrabbitLTC

Development/Code Analysis & Generationcode reviewRepomixstatic analysis

A custom MCP server that performs code reviews using Repomix and LLMs. It provides structured code reviews with specific issues and recommendations.

View on GitHub

Last updated: N/A

What is Code Review Server?

The Code Review Server is a custom MCP server designed to automate code reviews using Repomix for codebase flattening and Large Language Models (LLMs) for analysis. It provides structured feedback, identifying issues and suggesting improvements.

How to use Code Review Server?

  1. Clone the repository. 2. Install dependencies using npm install. 3. Build the server with npm run build. 4. Configure your LLM provider in the .env file. 5. Start the server using node build/index.js. 6. Use an MCP client to interact with the analyze_repo and code_review tools.

Key features of Code Review Server

  • Flattens codebases using Repomix

  • Analyzes code with Large Language Models

  • Provides structured code reviews with specific issues and recommendations

  • Supports multiple LLM providers (OpenAI, Anthropic, Gemini)

  • Handles chunking for large codebases

Use cases of Code Review Server

  • Automated code quality assessment

  • Identification of security vulnerabilities

  • Performance bottleneck analysis

  • Code maintainability improvement

  • Integration with MCP-compatible clients for streamlined workflows

FAQ from Code Review Server

What LLM providers are supported?

The server supports OpenAI, Anthropic, and Gemini.

How do I configure the LLM provider?

You can configure the LLM provider and API key in the .env file.

What is Repomix used for?

Repomix is used to flatten the codebase into a textual representation for analysis by the LLM.

What is the output format of the code review?

The code review is returned in a structured JSON format with a summary, issues, strengths, and recommendations.

How do I specify which files to review?

You can use the specificFiles parameter with the code_review tool or the --files option with the CLI tool.