🧩 Simple Comparer – OpenAI Model Evaluator

An interactive FastAPI + Tailwind + Chart.js web app to compare OpenAI models side-by-side. It lets you send the same prompt to two different models, visualize responses, and automatically evaluate them using an LLM-based or heuristic judge.

Key features:

• 🔹 Real-time comparison of OpenAI models (GPT-4o, GPT-4.1, GPT-3.5, etc.)
• 🔹 Automatic evaluation with structured metrics (Clarity, Task Fit, Structure, Safety, Correctness)
• 🔹 Topic extraction and shared-topic visualization via Venn diagrams
• 🔹 Compact radar, bar, and scatter charts for insights
• 🔹 Built with FastAPI, TailwindCSS, Jinja2, and Chart.js
• 🔹 CSV export of full comparison history

Ideal for: quick model comparisons, LLM evaluation demos, and educational projects.

---

⚙️ Quick start

git clone https://github.com/yourname/simple-comparer.git
cd simple-comparer
pip install -r requirements.txt
echo "OPENAI_API_KEY=sk-proj-..." > .env
uvicorn main:app --reload

Then open http://localhost:8000 and start comparing models.

---