AI Code Review Action – Automated PR Reviews Using AI (OpenAI, Gemini, ...)

quanlegia — Tue, 26 Aug 2025 05:06:24 +0000

Hey dev.to community!

I'm excited to announce that I've open-sourced AI-Powered Code Review Action – a GitHub Action that leverages state-of-the-art AI models like Gemini and OpenAI GPT-4 to automate your pull request code reviews.

Key Features

🔌 Pluggable Architecture: Easily add new/experimental AI providers.
🎯 Smart Filtering: Select files for review using flexible include/exclude glob patterns.
📦 Chunking Support: Handles large diffs by splitting them into chunks.
🎨 Customizable Rules: Add your own review rules and priorities.
💰 Cost Tracking: Monitor token usage and estimated API costs.
🔒 Secure: API keys are never logged; only diffs are sent to AI providers.
📊 Detailed Reports: In-depth review statistics and reporting.

Current Provider Support

Provider	Status	Model	Notes
Gemini	✅ Ready	gemini-1.5-flash	Fast, latest from Google
OpenAI	✅ Ready	gpt-4-turbo-preview	High quality, higher cost
HuggingFace	🚧 Soon	Various	Open source models planned

Quick Start

Create a new workflow at .github/workflows/ai-code-review.yml:

name: AI-Powered Code Review
on:
  pull_request:
    types: [opened, synchronize]
permissions:
  issues: write
  pull-requests: write
  contents: read
jobs:
  ai-review:
    runs-on: ubuntu-latest
    steps:
      - name: AI-Powered Code Review
        uses: legiaquan/code-review-action@v1.0.1
        with:
          provider: 'gemini'
          api_key: ${{ secrets.GEMINI_API_KEY }}

Easily configurable for more advanced usage!

How to Contribute 🚀

This is an early-stage project and all contributions are very welcome!

You can help by:

Adding support for new AI models/providers (open source or SaaS)
Improving tests and CI/CD
Suggesting or refining review rules
Fixing bugs, enhancing performance
Improving documentation, examples, and usability

👉 See the CONTRIBUTING.md for more details

👉 The code is written in TypeScript.

Example

👉 See the live example review by the bot on a demo PR: #21

Final thoughts

If you’re interested in AI-powered dev tools and automation, please star the repository, try integrating it with your projects, and especially send your pull requests and suggestions.

Thank you for reading and happy contributing!

AI-Powered Code Review Action

Refactor like a Pro: Enhance Your Code Quality and Maintainability! 🚀

quanlegia — Sun, 23 Mar 2025 05:28:15 +0000

🚀 Code Refactoring Techniques in Software Engineering 🔥

Refactoring is a crucial aspect of software development that enhances code structure while preserving its external behavior. It improves readability, maintainability, and scalability, ensuring long-term success for your projects.

🔎 In this in-depth guide, we explore:

✅ Understanding Refactoring – When and why should you refactor?

✅ Core Refactoring Principles – Embracing 🏗️ #DRY, 🎯 #KISS, ⚖️ #YAGNI, and other essential principles.

✅ Detecting Code Smells – Tackling 🛑 #LongMethod, 📏 #LargeClass, 🔁 #DuplicatedCode.

✅ Essential Refactoring Techniques – Applying 🛠️ Extract Method, ✍️ Rename Variable, and more.

✅ Advanced Approaches – Leveraging 🏗️ Replace Conditional with Polymorphism, 🎛️ Introduce Parameter Object.

✅ Refactoring in Software Architecture – Using Design Patterns to enhance system design.

✅ Tools for Refactoring – Harnessing SonarQube, ESLint, and smart IDE features.

✅ Hands-On Practice – Real-world exercises to sharpen your skills.

✅ 🎁 Bonus Content – Handling multilingual emails & diverse user types efficiently.

📖 Read the full guide here: 👉 Code Refactoring Techniques in Software Engineering

If you're committed to writing clean code or looking to elevate your refactoring skills, this guide is a must-read! 🚀

💬 What's the most challenging code smell you've encountered in your projects? Let’s discuss in the comments! 👇

DEV Community: quanlegia