PromptRefiner

A prompt optimization tool that analyzes, refines, and estimates costs for LLM prompts. Supports both OpenAI and Anthropic models.

What It Does

• Token Analysis — Counts tokens for your prompt across different models (GPT-4o, GPT-3.5-turbo, Claude 3.5 Sonnet, Claude 3 Haiku)
• Cost Estimation — Calculates input/output costs based on current model pricing
• Prompt Refinement — Uses LLM-powered analysis to rewrite your prompt for clarity, specificity, and efficiency
• Side-by-Side Comparison — View original vs. refined prompt with token count and cost differences
• Multi-Provider Support — Switch between OpenAI and Anthropic as the refinement engine

Tech Stack

Layer	Technology
Backend	Python 3.11+ · FastAPI · Uvicorn
Frontend	React 18 · Vite · TailwindCSS
Tokenization	tiktoken (OpenAI) · anthropic-tokenizer
LLM Providers	OpenAI API · Anthropic API
Testing	pytest (backend) · Vitest (frontend)

Project Structure

PromptRefining/
├── backend/
│   ├── app/
│   │   ├── main.py              # FastAPI application entry point
│   │   ├── api/
│   │   │   └── routes/
│   │   │       ├── analyze.py   # Token counting & cost estimation
│   │   │       └── refine.py    # Prompt refinement endpoint
│   │   ├── core/
│   │   │   ├── config.py        # Settings & environment variables
│   │   │   └── pricing.py       # Model pricing data
│   │   ├── services/
│   │   │   ├── tokenizer.py     # Token counting service
│   │   │   ├── cost.py          # Cost calculation service
│   │   │   └── refiner.py       # LLM-based prompt refinement
│   │   └── models/
│   │       ├── request.py       # Pydantic request models
│   │       └── response.py      # Pydantic response models
│   ├── tests/
│   ├── requirements.txt
│   └── .env.example
├── frontend/
│   ├── src/
│   │   ├── App.jsx
│   │   ├── components/
│   │   │   ├── PromptInput.jsx
│   │   │   ├── AnalysisPanel.jsx
│   │   │   ├── RefinedOutput.jsx
│   │   │   ├── CostComparison.jsx
│   │   │   └── ModelSelector.jsx
│   │   ├── services/
│   │   │   └── api.js
│   │   └── utils/
│   │       └── formatters.js
│   ├── package.json
│   └── vite.config.js
├── README.md
├── CONTRIBUTING.md
├── ENHANCEMENTS.md
├── CHANGELOG.md
├── CurrentApp.md
├── localsetup.md
└── LICENSE

Quick Start

# Clone
git clone https://github.com/your-username/PromptRefining.git
cd PromptRefining

# Backend
cd backend
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
cp .env.example .env  # Add your API keys
uvicorn app.main:app --reload

# Frontend (new terminal)
cd frontend
npm install
npm run dev

Open http://localhost:5173 in your browser.

API Keys Required

You need at least one of:

• OpenAI API Key — Get one here
• Anthropic API Key — Get one here

See localsetup.md for detailed setup instructions.

API Endpoints

Method	Endpoint	Description
POST	/api/analyze	Analyze token count and cost for a prompt
POST	/api/refine	Refine a prompt and return comparison
GET	/api/models	List supported models with pricing
GET	/api/health	Health check

Screenshots

Coming soon — UI is under development.

License

MIT — see LICENSE for details.