NorthStar is a full-stack web application that uses Anthropic Claude to help students find the right scholarships, decode what committees truly value, and generate tailored, high-impact application essays.
Built for the November 2025 Anthropic AI Hackathon.
Recommends the most relevant scholarships based on academic background, interests, activities, and achievements.
AI uncovers hidden evaluation criteria such as leadership, community service, financial need, resilience, and innovation.
Selectable narrative styles inspired by real winning essays:
- Community Builder
- Innovator
- Overcoming Adversity
- Academic Achiever
Claude generates a personalized scholarship essay draft aligned with each award’s priorities and the student’s authentic story.
| Layer | Technology |
|---|---|
| Frontend | React (JavaScript/TypeScript) |
| Backend | Python (FastAPI) |
| AI | Anthropic Claude API |
| Database | Supabase (PostgreSQL, optional) |
- Node.js (v16+) & npm
- Python 3.9+
- Anthropic API Key
- (Optional) Supabase project
git clone https://github.com/kikotc/NorthStar.git
cd NorthStar/backend
Create virtual environment:
bash
Copy code
python3 -m venv venv
source venv/bin/activate # Windows: venv\Scripts\activate
Install dependencies:
bash
Copy code
pip install -r requirements.txt
Create .env:
ini
Copy code
ANTHROPIC_API_KEY=<your_key>
SUPABASE_URL=<optional_supabase_url>
SUPABASE_ANON_KEY=<optional_supabase_key>
Run backend:
bash
Copy code
uvicorn main:app --reload
API Docs → http://localhost:8000/docs
⚛️ Frontend Setup (React)
bash
Copy code
cd ../frontend
npm install
npm start
Frontend → http://localhost:3000
🎮 Usage Guide
1. Create Student Profile
Enter academic background, extracurriculars, and personal experiences.
2. Find Matching Scholarships
NorthStar recommends opportunities using a personalized “fit score.”
3. View Scholarship Priorities
AI extracts weighted criteria (e.g., Leadership 40%, Financial Need 35%, Service 25%).
4. Generate Essay Draft
Select a persona → Claude generates a tailored essay focused on what the committee values.
5. Refine & Edit
Edit, regenerate variations, tweak tone, or request AI improvements.
6. Finalize & Submit
Export the essay and use it in your real scholarship application.
🧠 Data & AI Methodology
📚 Scholarship Dataset
Includes ~30 curated scholarships stored via Supabase or local JSON.
📝 Winner Essay Insights
Persona generation is derived from publicly available winning essays and student success stories.
🤖 Prompt Engineering
Prompts include:
Priority extraction
Persona-based essay drafting
Tone refinement and gap analysis
🔒 Ethics
No fabricated achievements
AI stays aligned with student-provided experiences
Essays remain authentic and transparent
📂 Project Structure
kotlin
Copy code
NorthStar/
├── backend/
│ ├── main.py
│ ├── prompts.py
│ ├── utils/
│ ├── data/
│ └── requirements.txt
├── frontend/
│ ├── public/
│ ├── src/
│ │ ├── components/
│ │ ├── pages/
│ │ ├── hooks/
│ │ ├── utils/
│ │ └── App.js
└── data/
├── scholarships_sample.json
├── sample_profiles.json
└── example_winners/
🌐 Deployment
🌩️ Frontend (Netlify / Vercel)
Build production bundle:
bash
Copy code
npm run build
Deploy the build/ directory.
Update API endpoint if backend is deployed separately.
🐳 Backend (Railway / Render / AWS / Heroku)
Example Heroku Procfile:
txt
Copy code
web: uvicorn main:app --workers 1 --port $PORT
Required environment variables:
nginx
Copy code
ANTHROPIC_API_KEY
SUPABASE_URL
SUPABASE_ANON_KEY
🗄️ Supabase (Optional)
Import scholarship dataset
Configure RLS for public read-only
Add URL & anon key in backend .env
🔐 CORS
Configure FastAPI CORS settings to allow your deployed frontend domain.
⭐ Final Note
NorthStar eliminates confusion around scholarships by helping students:
Identify opportunities that truly fit
Understand what committees prioritize
Craft essays that resonate and stand out
With AI-powered insights and tailored drafting, NorthStar transforms an overwhelming process into a strategic advantage.