Avultor

AI response evaluation tool that compares two AI-generated answers using weighted scoring criteria.

What it does

Avultor enables rigorous evaluation of AI-generated responses by comparing two answers side-by-side using customizable weighted scoring criteria. Define your own evaluation dimensions, assign importance weights, and get structured assessments to determine which AI response better meets your quality standards. Ideal for benchmarking models, refining prompts, and ensuring output quality.

Tech Stack

• Next.js - React framework for production
• TypeScript - Type-safe development
• Supabase - Backend database and authentication
• shadcn/ui - Accessible component library

Getting Started

git clone https://github.com/1aday/evals.git
cd evals
npm install
npm run dev

Environment Variables

Create a .env.local file with the following:

SUPABASE_URL=your_supabase_url
SUPABASE_ANON_KEY=your_supabase_anon_key

---

Built by @1aday