Refyne — The Evolving AI Engineer

Refyne analyzes repositories to identify architectural bottlenecks, summarize structural patterns, and produce focused refactor recommendations. It pairs an Express-based analysis backend with a React + Vite front end so teams can explore issues and act on prioritized improvements.

Table of contents

• What it does
• Repository layout
• Technology stack
• Prerequisites
• Local setup & quick start
• Configuration
• Development workflow
• Build & deployment
• Contributing
• License

What it does

Refyne provides:

• Structural analysis of codebases to surface architectural hotspots.
• High-level summaries of module and dependency structure.
• Actionable suggestions for refactoring and improving maintainability.
• A UI for exploring historical analyses and recommendations.

Repository layout

Refyne/
├── server/         # Express backend, analysis pipeline, API endpoints
├── client/         # React + Vite front end (UI components, styles)
├── data/           # Persistent analysis results (history.json)
├── utils/          # Shared analysis helpers and utilities
├── README.md       # This file
└── LICENSE         # Project license (MIT)

Technology stack

• Backend: Node.js, Express
• Frontend: React, Vite
• Styling: Tailwind CSS (present in client)
• Data store: JSON files under data/ (simple persistent history)

Prerequisites

• Node.js (LTS recommended)
• npm (included with Node.js)

Local setup & quick start

1. Install and run the backend:

cd server
npm install
npm run dev

2. Install and run the frontend:

cd ../client
npm install
npm run dev

Open http://localhost:5173 to access the UI. The backend defaults to port 5000 and the client runs on 5173 when using the Vite dev server.

Configuration

• Copy server/.env.example to server/.env and provide required values.
• Important environment variables:
  • GEMINI_API_KEY — API key for Gemini (used by AI-powered recommendation features).
  • PORT — backend port (defaults to 5000 if unset).

If GEMINI_API_KEY is not set, AI integrations will be skipped and the local analysis features remain available.

Development workflow

• Make UI changes in client/src/components/ and styles in client/src/styles/.
• Update analysis code in server/ and utils/.
• Analysis results and history are written to data/history.json for review.

Testing: there is no test suite included by default. Adding unit and integration tests is recommended prior to major changes.

Build & deployment

1. Build frontend for production:

cd client
npm run build

2. Serve the build output from a static host or integrate it with the Express backend for a single deployment artifact.

Recommended options:

• Host static files on platforms such as Vercel, Netlify, or an object storage + CDN.
• Or copy client/dist into the backend's static assets and serve via Express behind a reverse proxy.

Contributing

Contributions are welcome. Suggested workflow:

1. Fork the repository and create a feature branch: git checkout -b feature/your-feature.
2. Implement changes and include tests where applicable.
3. Open a pull request with a clear description and motivation for the change.

Guidelines:

• Keep pull requests focused and small when possible.
• Update README.md and include usage examples for new features.

If you'd like help scoping a contribution, open an issue describing the goal and maintainers can help plan the work.

License

This repository is licensed under the MIT License — see the LICENSE file for details.

Contact

Open an issue for questions, bug reports, or feature requests.