CogniCode Agent 🧠

Multi-agent AI system for real-time code analysis, refactoring, and test generation

CogniCode Agent is a privacy-first, AI-powered development tool that provides real-time code analysis, intelligent refactoring suggestions, and automated unit test generation. Built for Hackathon 2025 with a focus on developer productivity and code quality.

✨ Features

🔍 Real-time Code Analysis: Instant bug detection, style checks, and performance insights
⚡ Intelligent Refactoring: AI-powered suggestions to improve code quality and performance
🧪 Automated Test Generation: Comprehensive unit test creation with edge cases
🔒 Privacy-First: All AI processing runs locally - your code never leaves your machine
🌐 Multi-language Support: JavaScript, TypeScript, Python, Java, and more
📊 Code Metrics: Complexity analysis, maintainability scores, and quality insights
🎨 Beautiful UI: Modern, responsive design with real-time updates

🚀 Quick Start

Prerequisites

Node.js 16+ and npm
Python 3.8+ and pip
4GB+ RAM (8GB+ recommended for AI models)
5GB+ free disk space for AI models

Automated Setup

# Clone the repository
git clone https://github.com/Edmon02/cognicode-agent.git
cd cognicode-agent

# Run the setup script
chmod +x setup.sh
./setup.sh

# Start the full application
./run-dev.sh

Manual Setup

Click to expand manual setup instructions

Frontend Setup

npm install

Backend Setup

cd server
python3 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt

Download AI Models (Optional)

cd server
python scripts/download_models.py

Start Services

# Terminal 1 - Backend
cd server && source venv/bin/activate && python app.py

# Terminal 2 - Frontend  
npm run dev

📖 Documentation

📚 Complete Documentation - Comprehensive guides and references

Quick Links

🚀 Installation Guide - Detailed setup instructions
📖 Basic Usage - Getting started guide
🎯 Features Overview - All capabilities explained
🏗️ Architecture - Technical architecture
📡 API Reference - Complete API documentation
🎬 Demo Guide - Professional demo materials

📖 Quick Usage

Open the Application: Navigate to http://localhost:3000
Write/Paste Code: Use the Monaco editor to input your code
Select Language: Choose from JavaScript, TypeScript, Python, Java, etc.
Analyze Code: Click "Analyze" to run AI-powered analysis
Review Results: Check the Analysis, Refactor, and Tests tabs
Apply Suggestions: Click "Apply" on refactoring suggestions
Download Tests: Export generated unit tests

👉 Detailed Usage Guide

Example Workflow

// Paste this code and click "Analyze"
function fibonacci(n) {
    if (n <= 1) return n;
    return fibonacci(n - 1) + fibonacci(n - 2);
}

console.log(fibonacci(10));

The AI will detect:

⚠️ Performance issue (exponential time complexity)
💡 Refactoring suggestion (memoization)
🧪 Generated unit tests with edge cases

🏗️ Architecture

Multi-Agent System

┌─────────────────┐    ┌────────────────────┐    ┌─────────────────┐
│   Linter Agent  │    │ Refactor Agent     │    │ TestGen Agent   │
│                 │    │                    │    │                 │
│ • Bug Detection │    │ • Code Optimization│    │ • Unit Tests    │
│ • Style Analysis│    │ • Pattern Improve. │    │ • Edge Cases    │
│ • Security Check│    │ • Performance Tune │    │ • Integration   │
└─────────────────┘    └────────────────────┘    └─────────────────┘
         │                       │                       │
         └───────────────────────┼───────────────────────┘
                                 │
                    ┌─────────────────┐
                    │  Code Service   │
                    │                 │
                    │ • Result        │
                    │   Processing    │
                    │ • Caching       │
                    │ • Coordination  │
                    └─────────────────┘

Tech Stack

Frontend

Next.js 13+ (App Router)
React 18 with TypeScript
Tailwind CSS + shadcn/ui
Monaco Editor
Socket.IO Client
Framer Motion

Backend

Flask + Flask-SocketIO
PyTorch + Transformers
CodeBERT (Analysis)
CodeT5 (Refactoring)
Multi-agent Architecture

AI Models

microsoft/codebert-base - Code analysis and bug detection
Salesforce/codet5-small - Code generation and refactoring
Local execution for privacy

🔧 Configuration

Environment Variables

Frontend (.env.local)

NEXT_PUBLIC_BACKEND_URL=http://localhost:5000
NEXT_PUBLIC_APP_NAME=CogniCode Agent

Backend (server/.env)

FLASK_ENV=development
PORT=5000
MODELS_PATH=./models
USE_DEMO_MODE=true
LOG_LEVEL=INFO

Model Configuration

Models are automatically downloaded to server/models/. For custom models:

# server/agents/base_agent.py
CUSTOM_MODELS = {
    'linter': 'your-custom/codebert-model',
    'refactor': 'your-custom/codet5-model',
    'testgen': 'your-custom/test-model'
}

🚀 Deployment

Local Development

./run-dev.sh

Production with Docker

# Build and run with Docker Compose
docker-compose up --build

# Or build individual services
docker build -t cognicode-frontend .
docker build -t cognicode-backend ./server

Cloud Deployment

Frontend (Vercel)

# Deploy to Vercel
npm run build
vercel --prod

Backend (Docker/Cloud)

# Build production image
docker build -t cognicode-backend ./server

# Deploy to your preferred cloud provider
# (AWS ECS, Google Cloud Run, Azure Container Instances, etc.)

🧪 Testing

Frontend Tests

npm test
npm run test:watch

Backend Tests

cd server
source venv/bin/activate
pytest
pytest --coverage

Integration Tests

# Start both services and run E2E tests
npm run test:e2e

📊 Performance

Benchmarks

Operation	Time	Memory
Code Analysis	<500ms	~200MB
Refactor Generation	<1s	~300MB
Test Generation	<1.5s	~250MB
Model Loading	~10s	~1GB

Optimization Tips

CPU: Use quantized models for faster inference
Memory: Enable model caching and batch processing
Storage: Models require 2-5GB disk space
Network: All processing is local (no API calls)

🤝 Contributing

We welcome contributions! Please see our Contributing Guide for details.

Development Setup

# Fork and clone the repository
git clone https://github.com/Edmon02/cognicode-agent.git

# Create a feature branch
git checkout -b feature/amazing-feature

# Make your changes and test
npm test
cd server && pytest

# Commit and push
git commit -m "Add amazing feature"
git push origin feature/amazing-feature

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🏆 Hackathon 2025

CogniCode Agent was built for Hackathon 2025, targeting these challenges:

🤖 AI Innovation: Multi-agent architecture with local AI models
⚡ Real-time Applications: WebSocket-based live code analysis
🔒 Privacy-First Solutions: All processing runs locally
🛠️ Developer Tools: Productivity-focused development experience

Awards Targeting

Best Use of AI
Best Solo Project
Most Innovative Solution
Best Developer Tool

🔗 Links

Demo: https://cognicode-agent.vercel.app
GitHub: https://github.com/Edmon02/cognicode-agent
Author: Edmon02
Portfolio: https://github.com/Edmon02

📞 Support

📧 Email: Create an issue
💬 Discussions: GitHub Discussions
🐛 Bug Reports: GitHub Issues

Built with ❤️ for developers by developers

⭐ Star this repo | 🍴 Fork it | 📝 Contribute

Name		Name	Last commit message	Last commit date
Latest commit History 14 Commits
app		app
components		components
docs		docs
hooks		hooks
lib		lib
server		server
types		types
.env.example		.env.example
.gitignore		.gitignore
CONTRIBUTING.md		CONTRIBUTING.md
Dockerfile.frontend		Dockerfile.frontend
LICENSE		LICENSE
README.md		README.md
bun.lock		bun.lock
components.json		components.json
docker-compose.yml		docker-compose.yml
jest.config.js		jest.config.js
jest.setup.js		jest.setup.js
next-env.d.ts		next-env.d.ts
next.config.js		next.config.js
package-lock.json		package-lock.json
package.json		package.json
postcss.config.js		postcss.config.js
setup.sh		setup.sh
start-dev.sh		start-dev.sh
tailwind.config.ts		tailwind.config.ts
tsconfig.json		tsconfig.json

Folders and files

Latest commit

History

Repository files navigation