A sophisticated agentic AI system that accelerates developer onboarding through intelligent mentoring, personalized learning paths, and contextual task recommendations.
- 4 Specialized AI Agents: Knowledge, Guide, Mentor, and Task agents working in harmony
- Intelligent Code Analysis: Vector-powered search across your entire codebase
- Personalized Learning Paths: Adaptive guidance based on role and experience
- Real-time Mentoring: Contextual help and troubleshooting assistance
- Smart Task Recommendations: Curated starter tasks matched to skill level
- Modern Tech Stack: FastAPI backend with Next.js frontend
Experience ZeroDay's capabilities with synthetic data and mock integrations:
- Interactive chat with all 4 AI agents
- Sample learning paths and task suggestions
- Simulated code search and analysis
- Demo Slack integration workflows
Backend:
- FastAPI with async/await architecture
- ChromaDB vector database for semantic search
- OpenAI/Anthropic LLM integration
- Loguru for structured logging
Frontend:
- Next.js 14 with TypeScript
- Tailwind CSS for modern styling
- Lucide React icons
- Responsive design system
Infrastructure:
- SQLite for local development
- Docker-optional deployment
- Vercel-ready frontend
- Environment-based configuration
Python 3.9+
Node.js 18+git clone https://github.com/mukuldesai/ZeroDay
cd zeroday
python -m venv env
source env/bin/activate # Windows: env\Scripts\activate
pip install -r requirements.txt
python api/main.pycd frontend
npm install
npm run devVisit http://localhost:3000 to see ZeroDay in action.
# Ask the Knowledge Agent
POST /api/chat
{
"message": "How does the authentication system work?",
"agent_type": "knowledge"
}
# Get learning guidance
POST /api/chat
{
"message": "I want to learn React",
"agent_type": "guide"
}# Upload codebase for analysis
POST /api/upload
{
"file": "codebase.zip",
"data_type": "github_repo"
}ZeroDay uses a modular agent architecture where each AI agent specializes in different aspects of developer onboarding:
- Knowledge Agent: Searches and analyzes your codebase, documentation, and PRs
- Guide Agent: Creates personalized learning roadmaps and tracks progress
- Mentor Agent: Provides real-time help, debugging assistance, and guidance
- Task Agent: Recommends appropriate starter tasks based on skill level
Key environment variables:
OPENAI_API_KEY=your_key_here
ANTHROPIC_API_KEY=your_key_here
VECTOR_STORE_PATH=./data/vectorstore
UPLOAD_DIR=./uploadsPOST /api/chat- Main chat interfacePOST /api/upload- Data upload and processingGET /api/agents- List available agentsPOST /api/ask_mentor- Direct mentor queriesPOST /api/generate_plan- Create learning plansPOST /api/suggest_task- Get task recommendations
python api/main.py &
cd frontend && npm run dev- Deploy frontend to Vercel
- Deploy backend to Railway/Render
- Configure environment variables
- Update CORS settings
This is a portfolio project showcasing modern AI agent architecture and full-stack development skills. The codebase demonstrates:
- Clean FastAPI async patterns
- Modular agent design
- Vector database integration
- Modern React/TypeScript practices
- Professional deployment strategies
MIT License - feel free to use this code for learning and portfolio purposes.