LLM Council
A lightweight local web app that orchestrates a multi-model deliberation system. Queries are sent to multiple LLMs simultaneously via OpenRouter, models anonymously review each other's responses, and a designated Chairman LLM synthesizes perspectives into a final answer. Operates in three stages: First Opinions parallel responses, Review anonymous peer critique, and Final Response Chairman synthesis. Created as a "fun Saturday hack" by Andrej Karpathy to explore and evaluate multiple LLMs side by side.
- Radar Signals0