Skip to content

Nialled69/Note-gpt

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

18 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Note-GPT

An AI-powered simple notes app built with ❤️(and labour) using Next.js, React, TailwindCSS, TypeScript and ShadCN UI.

Capture your thoughts and turn them into intelligent summaries — fast, clean, and beautifully dark-mode ready.


🧪 Tech Stack

Layer Tech
Framework Next.js
Database Supabase
Auth Supabase + Google Cloud client
AI tool DeepSeek R1
Styling TailwindCSS + shadcn/ui
State handling React Query
Toasts UI Sonner

⚙️ Features

  • ✍️ Write and manage your personal notes
  • 🧠 Generate AI-powered summaries using powerful DeepSeek R1 LLM model
  • 🔐 Secure login with Supabase Auth (Google + regular Emails)
  • 🌗 Dark mode toggle (because... why not?)
  • 💬 Toast notifications via Sonner
  • ♻️ Auto-refreshing state with seamless handling of cookies and tokens
  • 🧼 TypeScript + ESLint + Prettier enforced

🚀 Getting Started

1. Cloning the repo

git clone https://github.com/Nialled69/Note-gpt.git
cd note-gpt

2. Installing the dependencies

npm i
# or
yarn i
#or
pnpm i

3. Settin up environment variables

Create a .env.local file in the root and add these:

NEXT_PUBLIC_SUPABASE_URL=<Your supabase project URL without quotes>
NEXT_PUBLIC_SUPABASE_ANON_KEY=<Your supabase project anon Key without quotes>
OPENROUTER_API_KEY=<Your LLM model API key without quotes>
NEXT_PUBLIC_SITE_URL=http://localhost:3000
NEXT_PUBLIC_SITE_NAME=Note-GPT

4. Run in localhost

npm run dev

Open http://localhost:3000 and ✨ start writing, editing and summarizing your notes!


🧠 AI Summarization

Powered by the blazing-fast DeepSeek R1 API.

  • Summary generation is triggered via a Summarize button on each note.
  • The summarizeNote server action sends note content to the DeepSeek R1 API.
  • Responses are parsed with full TypeScript safety and fallback error handling.
  • Toast messages via sonner display feedback for success, rate-limits, or unknown errors.
  • Server error parsing uses strict unknown type narrowing for clean logs and lint-free code.

🌍 Live Site

🔗 https://note-gpt-oo.vercel.app


👨‍💻 Author

Built with and late-night commits by Sarthik
Give a ⭐️ if you like it!