An AI-Native Full-Stack Todo System built using Spec-Driven Development (SDD), evolving from a simple CLI application into a modern distributed system with AI agents, authentication, backend APIs, and cloud deployment readiness.
This project demonstrates the progressive evolution of a real-world software system, starting from a simple CLI Todo app and expanding into a full-stack AI-powered application.
It focuses on how modern applications are built using:
- Spec-Driven Development (AI-assisted engineering workflow)
- Full-Stack architecture (Next.js + FastAPI)
- Secure authentication systems (JWT-based auth)
- AI Agents for natural language task management
- Cloud deployment preparation (future phases)
Instead of traditional development:
❌ Write Code → Debug → Refactor
✅ Write Spec → AI Generates → Iterate → System Evolves
This project represents a shift from developer as coder → developer as system architect.
- Next.js (App Router)
- Dashboard-based UI
- Authentication flow integration
- Modern responsive UI design
- FastAPI (Python)
- Middleware-based architecture
- JWT Authentication system
- RESTful API design
- MCP (Model Context Protocol) integration
- Neon Serverless PostgreSQL
- Stores:
- User authentication data (sign up / sign in)
- User-specific todos
- Task metadata and status tracking
- OpenAI Agents SDK
- Natural language task management
- AI-powered Todo operations:
- Add tasks
- Update tasks
- Delete tasks
- Query tasks via chat
- Secure JWT-based authentication
- User registration and login system
- Middleware-protected API routes
- Each user has isolated todo space
- Fully integrated with database persistence
An intelligent chatbot powered by OpenAI Agents SDK that allows users to manage todos using natural language.
- “Add a meeting with team tomorrow at 10 AM”
- “Mark my urgent tasks as complete”
- “Delete all completed tasks”
The AI:
- Understands intent
- Calls backend APIs
- Updates database in real time
- Returns contextual responses
- Clean separation of frontend and backend
- Secure middleware-protected API layer
- AI agent acts as orchestration layer
- Database-driven persistent state
- Scalable modular backend design
- Backend is currently deployed on Hugging Face Spaces
- Frontend runs independently with backend integration
- System is functional end-to-end in current phase
Ayesh Sidd AI Engineer | Agentic AI Developer | Full-Stack Developer
- LinkedIn: linkedin.com/in/aaish-siddiqui
- Email: hafizayeshsidd@gmail.com
Phase I → CLI Todo App (In-Memory)
Phase II → Backend + Database + Authentication System
Phase III → Full-Stack AI Todo System (Chat + Dashboard + Agents)
Phase IV → Cloud-Native Deployment (Upcoming)
Phase V → Kubernetes + Distributed AI System (Upcoming)