An intelligent code review platform that leverages local LLMs to analyze code quality, security, and performance issues in real-time.
Built in ~8 hours for the Haufe 2025 Hackathon.
- Local LLM Integration - Powered by Ollama (CodeLlama 7B) for privacy-first code analysis
- Smart Issue Detection - Automatically identifies security, performance, bugs, style, and documentation issues
- Custom Coding Guidelines - Add team-specific rules and conventions
- AI-Powered Suggestions - Get actionable fix recommendations with code examples
- Effort Estimation - Automatic calculation of time needed to address findings(based on LLM's view)
- Multi-Language Support - JavaScript, TypeScript, Python, Java, Go, Rust, C++
- Drag-and-Drop Upload - Easy file submission with automatic language detection
- Clean, Minimalist UI - Built with Angular + Tailwind CSS
Auth interface with user-friendly design
Overview of code reviews with AI analysis summary, issue breakdown, and effort estimation
Submit code via drag-and-drop or paste, with custom coding guidelines support
Detailed findings categorized by severity with suggestions and code fixes
Backend:
- NestJS
- Prisma ORM
- SQLite Database
- Ollama (Local LLM)
- JWT Authentication
Frontend:
- Angular
- TypeScript
- Tailwind CSS
- RxJS
- Node.js 18+
- Ollama installed with CodeLlama model
ollama pull codellama:7b
cd backendnpm installnpx prisma generatenpx prisma db pushnpm run start:dev
cd frontendnpm installng serve
- Upload Code - Paste or drag-and-drop your code file
- Set Guidelines (Optional) - Add custom coding rules to enforce
- AI Analysis - Local LLM analyzes the code
- Review Findings - Get categorized issues by severity
- Apply Fixes - Copy suggested code improvements with one click
- Local LLM integration (Ollama)
- Custom coding guidelines enforcement
- Effort estimation system
- Clean, intuitive UI/UX
- File upload with auto-detection
- Real-time code analysis
- Severity-based filtering
Built with ☕ during Haufe 2025 Hackathon