Instead of trusting marketing claims, I created a blind testing platform where you compare two anonymous AI responses and vote for the better one.
After 50 votes, you get personalized recommendations based on YOUR preferences — not some generic benchmark.
Key features: - Top AI models (GPT-5.2, Claude Opus 4.5, Gemini 3 Pro, Grok 4+) - Blind side-by-side comparison - Personal AI toolkit (unlocked after 50 votes) - Community-powered rankings by category - Expanding model selection based on demand
Tech stack: Next.js 16, Supabase, Tailwind, OpenRouter API, deployed on DigitalOcean.
Total investment: ~$97 (domain, hosting, tokens).
Would love your feedback — especially on: 1. Is 50 votes too high a barrier to unlock recommendations? 2. How would you monetize something like this? 3. What categories/prompts would you want to test?
Live at: https://llmatcher.com/