Kiro and Smart Planner ✅

Planner that help you to do anything, built with cutting-edge IDE - Kiro! 🚀

Comment

🌟 Inspiration

I wanted a planner that actually helps you move from idea to action—without sending data anywhere.

🤖 AI-Powered Development

This project was developed with significant assistance from Kiro IDE, an AI-powered spec-to-code development environment that helped:

  • 📋 Requirements Analysis - Generated comprehensive project requirements and user stories
  • 🏗️ Architecture Design - Created detailed system architecture and component design documents
  • ✅ Task Planning - Broke down development into structured, manageable tasks
  • 💻 Feature Development - Implemented most of the core application features including:
    • React 19 frontend with TypeScript and modern hooks
    • FastAPI backend with proper async patterns
    • Database models and API endpoints
    • AI integration with Ollama for task generation
    • Real-time chat interface and task management

The project was built from scratch with 99% made by Kiro.

I've decided to attend 2 hackathons same time:

OpenAI Open Model Hackathon

Code with Kiro Hackathon

Existing task tools are - either overcomplicated or could not provide guidance.

I combined a local gpt-oss-20b + local application + calendar to create a planner that feels like a pusher instead of a passive checklist and used KIro IDE as assistant.

🧠 What It Does

Planner turns natural language into steps and keeps them actionable:

  • Write your ideas→ get generated tasks
  • Local-first — no data leaves your machine
  • Smart status lifecycle
  • Frontend optimized for fast iteration: React 19 + Vite + React Query + Zustand

🏗️ How I Built It

Layer Tech Notes
IDE Kiro AI first IDE that helped to plan and develop application
UI React 19, TypeScript, Tailwind, Zustand, React Query Fast dev loop + clear separation of server vs client state
API FastAPI (async) Clean routing, dependency injection, OpenAPI docs
Data SQLAlchemy + Alembic Migrations + enum modeling
DB SQLite (dev) / PostgreSQL (ready) Simple local spin‑up; scalable path
AI Ollama local models Deterministic prompt shaping; privacy preserved
Calendar Google Calendar API Optional integration via service layer
Validation Pydantic (backend) + Zod (frontend) Dual runtime + compile‑time safety

🧩 Challenges I Ran Into

  • Kiro is always forgetting context
  • Connecting Ollama to API
  • How to work with old and newly created task lists
  • Handling errors on each step

🏆 Accomplishments I Proud Of

  • Fully local AI workflow — can work without internet
  • Clean UI - it's easy to start planning new goals
  • I will actually use it to generate tasks and add it to my Google Calendar

📚 What I Learned

  • How to create AI empowered full stack application
  • Prepare AI model response for API
  • Local-first application development

🚀 What's Next for "Planner That Guides You To Do Anything" ✅

  • Analyzing if goal is currently achievable
  • Learn previous goals and actually propose new ones

Made with ❤️ for builders who want AI superpowers without surrendering privacy.

Built With

Share this project:

Updates